CA2362771A1 - Genetic analysis - Google Patents

Abstract

A method is described for use in whole genome analysis. The method - termed inter-population perfectly matched duplex depletion - can overcome many of t he limitations of current approaches based upon SNPs and linkage disequilibrium within isolated populations. Inter-population perfectly matched duplex depletion isolates a fragment (or fragments) containing differences between the "affected" and "unaffected" populations or cells. A convenient method - terminal restriction site profiling arrays (TRSPAs) - is described for the analysis of such fragments. A totally diagnostic internal control DNA is als o described which allows both the extent and exact nature of any partial digestion to be unambiguously determined for inter-population perfectly matched duplex depletion or TRSPA restriction.

Description

GENETIC ANALYSIS
INTRODUCTION
Limitations of the current approaches There are a number of limitations to carrying out association studies using single nucleotide polymorphisms (SNPs) and linkage m disequilibrium within human populations (see Science, Vol 278, p1580, (1997) for a review of such methods). We have no control over recombination frequency around a given locus or over past human genetic crossing. Some mutations will be closely correlated with nearby SNPs and others will not.
1_5 The need for whole genome analysis With the SNP and linkage disequilibrium approach (and many others), markers are essentially used as a surrogate for sequencing - the more markers, the better. The logical endpoint of the above argument is to ~o look at every base in the human genome - and carry out what could be termed a whole genome association study. In essence, the sequence at every base would be determined for the genome of each member of a phenotypically 'affected' and a phenotypically 'unaffected' population.
Statistical correlations (associations) could then be drawn between ~s sequence differences and phenotype. Such associations would have future predictive values for the phenotype of interest, knowing the genotype and could be of great value in medicine and pharmacogenetics.
The current invention selectively enriches for DNA fragments that determine phenotype in the 'affected' population and thus makes the :~o prospect of carrying out whole genome association studies for humans and other species a very real possibility.

Definition of terms used with the current invention Within the scope of the current invention, the individuals chosen for whole genome analysis may be human, animal or plant and they may be eukaryotic, prokaryotic or archaebacterial in origin.
The terms 'affected' and 'unaffected' are used without limitation i~ arder to categorise individuals into two groups - those that possess a defined phenotype of interest ('affected' individuals) and those that do not possess the phenotype of interest ('unaffected' individuals).
The phenotype common to the 'affected' individuals may be either io beneficial (e.g. for these individuals, a particular pharmaceutical entity might show high efficacy in a phase II clinical trial) or detrimental (e.g.
for these individuals, a particular pharmaceutical entity might show adverse toxicology in a phase I clinical trial).
The 'affected' population may comprise one or more is individuals and the 'unaffected' population may similarly comprise one or more individuals according to the particular embodiment of the invention (see below).
The term DNA is used throughout for simplicity. Within the scope of the current invention, the term DNA may equally well apply to all ?o or part of the haploid, diploid or polyploid genomic DNA content of one or more germ line or somatic cell(s). The DNA may be extracted from cells taken directly from the individual(s), the DNA may be extracted from cells cultured or immortalised from the individuals) or the DNA may be prepared from a library of clones - with inserts derived from the individuals) and 2s propagated in some appropriate host and cloning vector system. For the particular case wherein the term DNA refers to the expressed part of the haploid, diploid or polyploid genomic DNA content of one or more somatic cells and the DNA is prepared from a library of clones - with inserts derived from the individuals) and propagated in some appropriate host and cloning o vector system, a cDNA library (normalised or otherwise) may be used.

In the current invention. DNA is compared in fragmented form. Fragmentation can be performed after DNA extraction, prior to cloning and/or after cloning. Restriction enzyme digestion is the preferred method for such fragmentation - though other methods (e.g. shearing or sonication) will be obvious to those skilled in the art.
For the particular case wherein the DNA is prepared from a library of clones (either genomic clones or cDNA clones) - with inserts derived from the individuals) and propagated in some appropriate host and cloning vector system and wherein restriction enzyme fragmentation is m used prior to cloning, polymerase chain reaction amplification can be used to prepare the DNA for comparison in fragmented form. Priming sites within the vector sequence flanking the cloned restriction enzyme fragmented inserts may be usefully employed for one or more cycles of polymerase chain reaction amplification of the fragmented DNA of interest.
> > The primers used for polymerase chain reaction amplification of the fragmented DNA of interest could again be used after the phenotype-determining fragment enrichment process to 'rescue' and clone the enriched fragments.
Within the scope of the current invention, the terms ~o biotinylation and streptavidin capture are used both as an example and as the currently preferred embodiment for the invention. The streptavidin may be surface attached to inert particles (magnetic or otherwise) or to vessel walls (e.g. microtitre plate wells). The biotin may be introduced via a deoxynucleotide triphosphate analogue using a polymerase; by using a ~s biotin-conjugated primer and polymerase chain reaction amplification;
chemically or photochemically. The use of biotin and streptavidin is not a limitation for the invention. The invention could equally be used with other high affinity capture systems well known to those skilled in the art (e.g.
'his tag' introduction and metal ion affinity capture).
:~o Within the scope of the current invention, the term 'abnormal' - used with respect to the term 'normal' - is used without limitation in order 4 PCT/GB00/00916 =

to denote a somatic cell (or somatic cells) with a discernable phenotypic characteristic (or characteristics) arising from the acquisition or' a different somatic mutation (or set of somatic mutations) from that (or those) seen in the 'normai' counterpart. Cells will most usually be considered 'abnormal' with respect to their'normal' counterparts through the acquisition of a different somatic mutation (or set of somatic mutations) leading to one or more of the following phenotypic characteristics: altered marker gene expression, altered genomic organisation, growth under certain selective culture conditions, immortalised growth in culture, unrestrained growth in m vivo or in vitro, failure of normal apoptotic control mechanisms in vivo or in vitro, induction of neovascularisation, escape of cells across epithelium, migratory cell survival or metastasis.
Within the scope of the current invention, the term mismatch recognition protein is used without limitation to denote a protein of ~s eukaryotic, prokaryotic or archaebacterial origin capable of the selective recognition of (and binding to) a DNA duplex that is not perfectly matched along its entire length. Recognition of (and binding to) will be preferably for bases that are not engaged in correct Watson and Crick pairing and for small deletions or insertions. Many such proteins are known to those o skilled in the art. Prokaryotic and eukaryotic mutS homologues, phage T4 endonuclease VII, phage T7 endonuclease I and the plant enzyme CEL-1 are just some examples.
Inter-population perfectly matched duplex depletion ~s In the inter-population perfectly matched duplex depletion approach, we compare (in fragmented form) the pooled DNA of 'affected' individuals with the pooled DNA of 'unaffected' individuals (both from populations as outbred and otherwise similar to each other as possible).
We are only interested in those regions where differences occur between ~o 'affected' and 'unaffected' DNA molecules. For populations as above, the only prevalent sequence differences within the 'affected' pool (compared to WO 00/55364 PCT/GB00/00916 =

the 'unaffected' pool) should be somewhere within the genes) (using the term gene in its widest sense to include exons. introns and all associated upstream and downstream regulatory sequences) that actually determines) their common phenotype. This means that we are no longer tied into working with rare (and perhaps atypical) populations where there is high genetic homogeneity.
Pooling the DNA from entire phenotypically-defined populations massively reduces the amount of labour involved.
m DNA sequence variation in populations Nickerson et al, Nature Genetics. Vol 19, p233 (1998), sequenced 9.7 kb of the lipoprotein lipase gene from 71 individuals (24 African-Americans, 24 Europeans and 23 European-Americans). This gene is fairly typical (90 intron and 10 % exon - total size 30 kb with 10 exons).
~ s 88 sequence variants were found (i.e. one per 110 by on average). Most variations were found in non-coding sequence. 90 % of these were SNPs (60 % of which were transitions and 40 % were transversions). All of the SNPs were biallelic. 10 % of the sequence variants were insertions or deletions at repeat sequences.
~0 58 °o of the sequence variants were present in all three ethnic populations. Half of these were found in heterozygous form and half in homozygous form.
Nucleotide diversity (defined as the expected number of nucleotide differences per site between a random pair of chromosomes ~s drawn from the population) is 1/500 for DNA in general and 1/2,000 for coding sequence DNA. This means that, on average, any two DNA
fragments annealed together from such a population will contain a mismatch every 500 bases.
DNA sequence variants are therefore very common. They ~o are not, however, totally random - the variants that occur every 500 bases or so are limited; they are generally biallelic at just that single base. It is this fact that the inter-population perfectly matched duplex depletion approach selectively exploits.
Statistical analysis of the inter-population perfectly matched duplex depletion fragmentation process If the length of the DNA fragments is F bases and the average length between sequence differences between any two DNA
molecules is 500 bases, the probability that a hybrid duplex between any two random DNA fragments will contain no mismatches is given by to Pr(0) = eOFiSOO) and the probability that a hybrid duplex between any two random DNA
molecules will contain any mismatches is given by Pr(>_1) _ {1-(e-~F~500))~
Example values for Pr(0) and Pr(>_1 ) for different values of F
are given below F Pr(0)Pr(>_1 ) 10 0.980.02 0.960.04 50 0.900.10 100 0.820.18 200 0.670.33 300 0.550.45 400 0.450.55 500 0.370.63 600 0.300.70 700 0.250.75 800 0.200.80 900 0.170.83 1000 0.140.86 _7_ Example average restriction fragment sizes for DNA digestion with six 6 by cutters and up to four 4 by cutters Each 6 by cutter will cut DNA every 4.096 by on average and each 4 by cutter will cut DNA every 256 by on average.
For a given set of A 6 by cutters and B 4 by cutters (with no duplication of restriction enzyme cutting sequence), the average fragment length F will be {(A/4,096)+(B/256)}-, io Example values for F as A and B are varied are given in the following table A B average size (F) ~s We should note that the above situation assumes that none of the 4 by cutter recognition sites lie within any of the 6 by cutter recognition sites. If, for example, we have a 4 by cutter recognition site nested within a 6 by cutter recognition site (e.g. from the use of Mbo! and BamHl in the fragmentation), then we should reduce the value of A from 6 to 5 in the 2o above.
In general, if we have a given set of A enzymes that cut DNA
every a by on average, B enzymes that cut DNA every b by on average, ... , Z enzymes that cut DNA every z by on average (with no duplication of restriction enzyme cutting sequence), the average fragment length F will be ?s {(A/a)+(B/b)+ ... + (Z/z)}-, _g_ Example sets of 6 by cutters and 4 by cutters for fragmentation Example sets of 6 by cutters and 4 by cutters that contain panels of six 6 by cutters that are compatible with terminal restriction site profiling array (TRSPA) analysis (see below) are given in the following:
Example enzyme set 1 For restriction in M buffer + BSA
Number Enzyme Site M+BSA OptimumU / SupplierInactivate pl o C C/min activity 1 Mbol GATC 100 37 25 NEB 65/20 2 Haelll GGCC 100 37 50 NEB 80/20 3 ~Ylsel TTAA 100 37 I 20 NEB 65/20 ~

Number Enzyme Site M+BSA OptimumU / SupplierInactivate C pl C/min activity 1 BamHl GGATCC 100 37 100 NEB 80 / 20 2 BsrGl TGTACA 100 37 10 NEB 80 / 20 3 ~ HindlllAAGCTT 100 37 >40 APB 65 / 20 Ncol CCATGG 100 37 50 NEB 65 / 20 Spel ACTAGT 100 37 50 NEB 65 / 20 6 Aflll CTTAAG 100 37 10 APB 60 / 15 Example enzyme set 2 For restriction in M buffer + BSA
Number Enzyme Site M+BSA OptimumU / SupplierInactivate C pl C/min activity 1 Mbol GATC 100 37 25 NEB 65/20 2 Haelll GGCC 100 37 50 NEB 80/20 3 Msel TTAA 100 37 20 NEB 65/20 _g_ Number Enzyme Site M+BSA Optimum U ~' ul SupplierInactivate -o =C =-C/min activity 1 EcoRl GAATTC 100 37 >40 APB 65 / 20 2 BspHl TCATGA 100 37 i 0 I APB 65 / 20 3 Bglll AGATCT 75 37 >40 APB No 4 Xbal TCTAGA 100 37 100 ~ NEB 65 / 20 Acc651 GGTACC 75 37 10 NEB 65 / 20 6 ApaLl GTGCAC 100 37 10 NEB No Aspect (I) In one aspect the invention provides a method of providing a mixture of DNA fragments enriched in fragments that are characteristic of a phenotype of interest, by providing affected DNA in fragmented form and providing unaffected DNA in fragmented form, which method comprises:
a) mixing the fragments of the affected DNA and the fragments of the unaffected DNA under hybridising conditions;
b) recovering a mixture of hybrids that contain mismatches;
to c) recovering fragments of the affected DNA from the mixture of hybrids that contain mismatches;
and optionally repeating steps a), b) and c) one or more times.
~s Inter-population mismatch-containing duplex selection 'Affected' versus 'unaffected' (i.e. inter-population) mismatch-containing duplex selection can be achieved by attaching a mismatch-binding protein to a solid support (or using the mismatch-binding protein in solution followed by subsequent solid-phase capture), taking fragmented ~o and denatured 'affected' DNA and hybridising this to an excess of fragmented, denatured and biotinylated 'unaffected' DNA with ensuing capture of mismatch-containing duplex molecules. Releasing the mismatch-containing duplex molecules without denaturation, streptavidin capture and then release of the non-biotinylated strands will give only the desired species as shown below.
Method Fragment the 'affected' DNA. Fragment and derivatise the 'unaffected' DNA with biotin. Only DNA from this population will be streptavidin captured. Melt and anneal to give biotin '~~ biotin biotin Biotin biotin ' biotin Y
o Mismatch-binding protein select. Capture only the mismatch-containing duplexes. Release without denaturation to give ~ biotin biotin ,r-biotin Streptavidin capture to give biotin biotin biotin ,,~
Release the non-biotinylated strands to give ~ biotin Repeat as necessary.
Repetition of the above sequence of reactions will lead to inter-population perfectly matched duplex depletion.
io What will be purified ?
Inter-population mismatch-containing duplex selection as above ensures that all of the various phenotype-determining fragments (unique to the 'affected' population) are captured for subsequent analysis -oa but it also causes the co-purification of very many SNP-containing ('noise') fragments.
We now need to consider the fate of the various types of fragment (i.e. those that determine the phenotype and those that do not) as we carry out inter-population perfectly matched duplex depletion cycles as ~o above.

Recovery of 'affected' DNA molecules after streptavidin capture For a particular fragment. if we have X molecules of 'affected' DNA and Y molecules of 'unaffected' biotinylated DNA. after complete hybridisation, there will be a ratio of {Y/(X+Y)} streptavidin-capturable molecules to {X/(X+Y)} streptavidin-non-capturable molecules. We can thus manipulate the yield of streptavidin-capturable hybrids by varying X
and Y.
Example recovery and loss figures for various X and Y are shown in the following table io ratiorecovery Loss ~ _ {Y/(X+Y)}_ {X/(X+Y)}
(Y/X) 1 50% 50%

2 67 ~0 33 3 75% 25%

4 80 % 20 9 90% 10%

19 95% 5%

99 99% 1 999 99.9 % 0.7 After n cycles, the recovery for a phenotype-determining fragment will be given by {Y/(X+Y)}"
i s Loss of general SNP-containing ('noise') fragments during inter-population perfectly matched duplex depletion cycles If we anneal fragmented DNA molecules together and capture only the mismatch-containing duplexes, then n repetitions of such a process will reduce the original number of fragments to the following 2o fraction {1 _(e- ~F/5oo))}~'{Y/(X+Y)}"

The enrichment for phenotype-determining fragments over SNP-containing ('noise') fragments during inter-population perfectly matched duplex depletion cycles will therefore be given by J {1-(e (F~500))~-n Example figures for enrichment are given below with F = 50, 100, 200. 300. 400, 500, 600, 700, 800, 900 and 1,000 and n = 1, 2, 3, 4, 5, 6 ,7, 8, 9 and 10.
~o n=

50 11 1101,16012,194128,135 1.E+071.E+082.E+09 2.E+10 1,346,489 100 6 30 168 926 5,110 155,500857,8414.732,4153.E+07 28,187 200 3 9 28 85 257 779 2.362 7,16621,735 65,929 300 2 5 11 24 53 119 263 582 1,291 2,860 1.0001 1 2 2 2 2 3 3 4 4 Loss of specific SNP-containing ('noise') fragments during inter-population perfectly matched duplex depletion cycles The non-polymorphic and SNP-containing ('noise') fragments ~s will be depleted as described above.
Not all fragments will, however, be depleted at the same rate.
An individual SNP-containing ('noise') fragment will be depleted with every cycle of inter-population perfectly matched duplex depletion as outlined below.
zo Let us assume that there are two alleles for the polymorphism within a particular fragment. Let these be called P and Q. Let p be the fraction of the P allele in the (outbred) 'affected' and 'unaffected' populations and let q be the fraction of the Q allele. Let (p+q)=1. so that q=( 1 _P).
After denaturation and annealing, the four possible events are PP, PQ, QP and QQ. PP and QQ will form perfectly matched duplexes and will therefore be lost - whereas PQ and QP will form mismatch-containing duplexes and will consequently be recovered.
After one cycle, the fraction of recovered molecules will be l o (2Pq~(p2+2pq+q2)) _ (2pq~((p+q)2)) = 2pq = 2p(1-p) Hence if we start out with M molecules of DNA from the population. there will be 2Mpq molecules remaining after the first round of ~o inter-population mismatch-containing duplex selection. Let us denote the number of molecules entering the second round of inter-population mismatch-containing duplex selection by M', where M' = 2Mpq and the fraction of lost molecules will be ~(P2+q2)~(p2+2pq+q2)) ~o = {(p2+q2)~((p+q)2)}

= P2+q2 = P2+f 1-2P+P2) a = 1-2p+2p2 = 1-2P(1-P) =1-2pq The fractional loss of P-allelic molecules will be p' and the fractional loss of Q-allelic molecules will be q2 ( _ (1-p)2).
If we start out with M molecules of DNA from the population, there wilt be Mp of the P-allelic molecules and Mq = M(1-p) of the Q-allelic is molecules before inter-population mismatch-containing duplex selection.
After inter-population mismatch-containing duplex selection, there will therefore be Mp-Mp2 = Mp(1-p) = Mpq of the P-allelic molecules and Mq-Mq2 = Mq(1-q) = Mqp of the Q-allelic molecules. In other words, after the first round of mismatch-containing duplex selection, there will be the same ~o number of P-allelic molecules as Q-allelic molecules.
We can define new allelic frequencies p' and q' as follows p'=q'=0.5 2s If we now perform a second round of inter-population mismatch-containing duplex selection, we start out with M' molecules of DNA from the first round. There will be M'p' of the P-allelic molecules and M'q' of the Q-allelic molecules before inter-population mismatch-containing duplex selection.
~o After inter-population mismatch-containing duplex selection, there will therefore be M'p'-Mpp' = M'p'(1-p) = M'p'q of the P-allelic molecules and M'q'-Mqq' = M'q'(1-q) = M'q'p of the Q-allelic molecules.
The total number of molecules (M") will be M'p'q + M'q'p = M'(p'q-q'p) but p'=q'=0.5 to hence M" = 0.5*M'(p+q) but (p+q) = 1 so M" _ (M'/2) In other words, after the second round of inter-population mismatch-containing duplex selection, there will again be the same number of P-allelic molecules as Q-allelic molecules. Thus the new allelic frequencies p" and q" remain as previously p" = q'' = 0.5.
~o A pattern now emerges. After the first cycle of inter-population mismatch-containing duplex selection, the allelic frequencies are equalled and the number of molecules is reduced to 2Mpq (Mpq of each allelic molecule). Thereafter, every cycle halves the number of molecules and keeps both alleles at the same frequency.
Consequently, after n cycles, the number of P-allelic molecules will be Mpq(0.5)"-' ~o and the number of Q allelic molecules will be Mqp(0.5)"-' both are, of course, equal.
If we now take the capture yield (see above) into consideration, the SNP-containing ('noise') fragment yield will be given by 2Mpq(~.5)"-1 ~{Y/(X+Y)}"
where both allelic variants are deemed to be captured 'noise'.
Polymorphisms that interfere with the pattern of restriction digestion For both the loss of general and specific SNP-containing ('noise') fragments during inter-population perfectly matched duplex depletion cycles (described above) and where the SNP interferes with the pattern of restriction digestion, if the mismatch-binding protein also binds to duplex molecules with unequal lengths (e.g. from inter-population annealing around a site of restriction site polymorphism), then the above analysis still holds (with perfectly matched duplex being replaced by equal length duplex and mismatch-containing duplex being replaced by unequal ~o partner-length duplex).
In the rare cases where a restriction site is lost due to a sequence change that actually determines the phenotype of interest, a double-length fragment will be obtained. This will give rise to a double terminal restriction site profiling array (TRSPA) signature (see below).
Multiple isolates of the particular double signature will be indicative of an association between the fragment and the phenotype of interest.

Further 'kinetic' enrichment to enhance the selective removal of SNP-containing 'noise' from the pool of phenotype determining fragments After multiple cycles of enrichment by the above procedure, the enriched DNA pool should contain many copies of all phenotype-determining fragments but also low numbers of copies of many different phenotype non-determining fragments. The total number of 'noise' fragments may exceed the number of phenotype determining fragments, despite the number of each individual 'noise' species being very small. The 'noise' fragments would therefore increase the number of probes required »> for TRSPA analysis before a pattern emerges. To largely eliminate this problem, a further kinetic enrichment procedure is used. Either one or both of strategies A and B below can be employed to achieve 'kinetic' enrichment.
~s Strategy A - Subtraction of the enriched DNA from inter-population mismatch containing duplex depletion The enriched fragment pool from inter-population mismatch containing duplex depletion is rapidly self-hybridised - enabling the common phenotype-determining fragments to form perfectly matched ~c> duplexes with greater efficiency than the rare 'noise' fragments.
Selection for perfectly matched duplexes then yields a selectively further enriched pool of fragments. Multiple cycles of subtraction could be carried out if necessary.
~s Strategy B - Hybridisation of the enriched DNA from inter-population mismatch containing duplex depletion against the 'affected' DNA pool The enriched fragment pool from inter-population mismatch containing duplex depletion is then hybridised to an excess of biotinylated DNA from the 'affected' pool. This allows the common phenotype-~o determining fragments to form perfectly matched duplexes with greater efficiency than the rare 'noise' fragments. Selection for perfectly matched duplexes followed by streptavidin capture and denaturation to release the non-biotinylated strands then yields a further enriched pool of fragments.
Multiple such 'affected' pool back-hybridisations could be carried out if necessary.
Extension of the invention to the case of single phenotypically 'affected' individuals within populations where the distinction between 'affected' and 'unaffected' is clear The above has described inter-population perfectly matched ~c~ duplex depletion between non-biotinylated DNA fragments from an 'affected' population and biotinylated DNA fragments from an 'unaffected' population. Provided the 'unaffected' population is sufficiently complex that it contains all the non-phenotype-determining sequence variants found in a single 'affected' individual, then inter-population perfectly matched duplex > > depletion should be possible for single phenotypically 'affected' individuals against an 'unaffected' population where the distinction between 'affected' and 'unaffected' is clear. The latter proviso is needed in order to ensure that a small number of misdiagnosed 'affected' individuals in the 'unaffected' population do not cause the removal of phenotype-determining ~o fragments during inter-population perfectly matched duplex molecular depletion.
Extension of the invention to the case of disease gene identification in cases where novel phenotype-determining mutations arise ''s spontaneously within a family Except for a small number of sequence changes, each of us contains DNA sequence derived from our parents - our individuality resulting from precisely which parental alleles we receive. If one of the above small number of sequence changes results in a change in ~o phenotype, then we can use inter-population perfectly matched duplex depletion to enrich for fragments encoding this change in phenotype.

If we take 'unaffected total ancestral' cells (by which we mean cells derived from a complete set of 'unaffected' ancestors - e.g. both parents, or mother plus two paternal grandparents, or father plus two maternal grandparents. or two maternal grandparents and two paternal grandparents etc. ) as the source of our biotinylated fragments and cells from an 'affected' descendent as the source of our non-biotinylated fragments, any fragments that have acquired phenotype-determining sequence changes between 'unaffected' ancestral generations and the 'affected' descendent generation will be unable to form perfectly matched m duplexes with the biotinylated 'unaffected total ancestral' fragments.
Successive cycles of such inter-population perfectly matched duplex depletion will thus lead to the enrichment of fragments carrying all such sequence - the degree of enrichment per cycle being as described below.
~ a Statistical analysis Let us assume that equal numbers of fragments are used from each of the 'unaffected' ancestors. Let the number of such ancestors be A.
{1/A} of the annealings in the inter-population perfectly ~o matched duplex depletion will be self-against-ancestral-transmitted alleles -statistically equivalent to self-against-self inter-population perfectly matched duplex depletion (see below).
{(A-1 )/A) of the annealings in the inter-population perfectly matched duplex depletion will be self-against-ancestral-nontransmitted ~s alleles - statistically equivalent to inter-population perfectly matched duplex depletion between unrelated individuals.
A {1/A} {(A-1)/A}
2 1 /2 'h 4 1 /4 3/a From the data of Nickerson et al, a DNA sequence variation between unrelated individuals should occur every 500 base pairs (it is this figure that we use for inter-population perfectly matched duplex depletion with self-against-ancestral-nontransmitted alleles and inter-population perfectly matched duplex depletion between unrelated individuals). In addition, a given individual should be heterozygous once every 573 base pairs (this figure is used for inter-population perfectly matched duplex depletion with self-against-ancestral-transmitted alleles and self-against-self inter-population perfectly matched duplex molecular depletion). Inter-»> population perfectly matched duplex depletion against the transmitted alleles and the nontransmitted alleles will now be considered separately.
Self-against-ancestral-transmitted alleles inter-population perfectly matched duplex depletion is If we anneal the two complementary strands for a fragment using DNA from an 'affected' descendent and DNA containing the transmitted alleles of 'unaffected' ancestors, then {1-e~-F~5'3~} of fragments will contain one or more site of heterozygosity. {1/A} of the annealings will be of this type. For such annealings, where a site of heterozygosity is ~o present, the probability of obtaining a mismatch-containing duplex between a biotinylated fragment and a non-biotinylated fragment containing the site of heterozygosity is 0.5.
Self-against-ancestral-nontransmitted alleles inter-population ~s perfectly matched duplex depletion If we anneal the two complementary strands for a fragment using DNA from an 'affected' descendent and DNA containing the nontransmitted alleles of 'unaffected' ancestors, then {1-e~~F~5oo>} of fragments will contain one or more site of DNA sequence variation. {(A-~0 1 )/A} annealings will be of this type.

WO 00/55364 PCT/GB00/00916 =

Phenotype-determining fragment enrichment The fraction of fragments carried through the first cycle of inter-population perfectly matched duplex depletion will therefore be ([0.5~{1 /A}.{ 1-e~-F'S~s~}~+[{(A-1 )/A}-{ 1-e~-F~soo>}~)'{Y/(X+Y)}
where {Y/(X+Y)} represents the yield of streptavidin-capturable molecules.
Hence n repetitions of such a process will reduce the original »~ number of fragments to the following fraction ([0.5~{1 /A}-{ 1-e~~F'S~s>}~+[{(A-1 )/A}'{1-e~-F~SOO~}~)"w{Y/(X+Y)}"
The enrichment for phenotype-determining fragments over is SNP-containing ('noise') fragments during inter-population perfectly matched duplex depletion cycles will therefore be given by ([0.5~ { 1 /A}~{ 1-e~-F!5~3>}]+[{(A-1 )/A}~{ 1-e~-F~500)}~)-n ~o Example figures for enrichment are given for A = 2, 3 and 4 below with F = 50, 100, 200, 300, 400, 500, 600, 700, 800, 900 and 1,000 andn=1,2,3,4,5,6,7,8,9and10.

For A=2 n=

50 15 2133,115145.494664.4279,703,6941.ET08 2.E+093.E+104.E+11 100 8 59 448 3.430 26.250200.8871,537.3671.E+079.E+077.E+08 200 4 18 74 309 1.2965.434 22.782 95.524400.5201,679,340 300 3 9 28 87 265 811 2,475 7,558 23.07870,468 400 2 6 16 39 97 241 600 1,497 3.7349,313 500 2 5 10 22 48 103 224 485 1,0512.277 [

I ~

800 2 3 5 8 ~ 14 24 42 71 121 205 1,0002 2 4 6 9 15 23 36 57 89 For A=3 n=

50 13 1672.15427,823359,3404,640,9416.E+078.E+08 1.E+101.E+11 100 7 46 311 2,104 14,25196,516 653.6824,427,2413.E+072.E+08 -200 4 14 51 191 709 2,633 9,786 36,365 135.135502,173 300 3 7 20 54 146 396 1.073 2,909 7.882 21,361 400 2 5 11 24 53 119 263 582 1.290 2,859 1,0001 2 3 4 5 8 11 15 21 29 For A=4 n=
~

50 12 1491,82622.316272,7593.333,7714.E+075.E+08 6.E+097.E+10 100 6 I 264 1.690 10,83469.465 445,3732.855.5122.E+071.E+08 200 4 12 44 154 540 1.902 6.696 23,570 82,967292.045 300 3 7 17 44 112 287 737 1.894 4,86412,494 400 2 4 9 20 41 86 181 381 800 1,681 ' I

1.0001 2 2 3 ~ 4 6 ~ 7 ~ 10 ~ 13 17 ~ ~ ~ ~ ~

Extension of the invention to the case of fully comprehensive 'abnormal' cell mutational profiling within an individual If we now take 'normal' cells as the source of our biotinylated fragments and 'abnormal' cells from the same individual as the source of our non-biotinylated fragments, any fragments that have acquired sequence changes on the way to becoming 'abnormal' will be unable to io form perfectly matched duplexes with the biotinylated 'normal' fragments.
Successive cycles of such inter-population perfectly matched duplex depletion will thus lead to the enrichment of fragments carrying all those sequence differences between the 'normal' cells and the 'abnormal' cells -the degree of enrichment per cycle being as described below.
~s From the data of Nickerson et al, a given individual should be heterozygous about once every 573 base pairs.
If we anneal the two complementary strands for a fragment using DNA from 'abnormal' cells and DNA from 'normal' cells, then {~_e~-Fis73>} of fragments will contain one or more site of heterozygosity.
2o For a heterozygous site, p and q are both 0.5. The probability of obtaining a perfectly matched duplex between a biotinylated fragment WO 00/55364 PCT/GB00/00916 =

and a non-biotinylated fragment containing the site is 0.5. Similarly, the probability of obtaining a mismatch-containing duplex between a biotinylated fragment and a non-biotinylated fragment containing the site is also 0.5.
The fraction of fragments carried through the first cycle of inter-population perfectly matched duplex depletion will therefore be 0.5~{ 1-e~-F~573)} ,{Y/(X+Y)}
0o hence n repetitions of such a process will reduce the original number of fragments to the following fraction (0.5~{1-e~-F~573)}~ n.{Y/(X+Y)}n )~ The enrichment for phenotype-determining fragments over SNP-containing ('noise') fragments during inter-population perfectly matched duplex depletion cycles will therefore be given by (0.5~{1-e~ F~573))~-n ?0 Example figures for enrichment are given below with F = 50, 100. 200, 300, 400, 500, 600, 700, 800, 900 and 1,000 and n = 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10.

WO 00/55364 PCT/GB00/00916 =

n=

I

50 24 57313.711328.1717,854.6272.E+084.E+091.E-113.E+12 6.E+13 ' 100 12 1561.94824.329303,8443.794,7315.E+076.E+087.E+09 9.E+10 200 7 46 313 2.12314.412 97.830664.0764.507.7843.E+07 2.E+OS
I

300 5 24 118 580 2.845 13.95868,490336.0721,649,0598,091,708 I

400 4 16 63 251 999 3,977 15,83163.012250.815998,346 500 3 12 41 139 479 1.644 5.650 19,41166,688 229,114 600 3 9 29 90 278 856 2,638 8,128 25,047 77,179 700 3 8 23 65 183 520 1.475 4,183 11,862 33,639 800 3 7 19 50 133 353 937 2,491 6,621 17,599 I I

900 3 6 16 41 103 259 654 1.652 4.171 10,532 1,0002 6 14 34 84 202 490 1,188 2,880 6,978 This approach will enrich for fragments containing all sequence differences within the 'abnormal' cells - no prior knowledge of the s genes (oncogenes, tumour suppressor genes etc. ) that may need to be investigated is required.
Having thus isolated a fragment (or fragments) determining differences between the 'affected' and 'unaffected' populations, we can now proceed to their analysis.

Aspect (II) In another aspect the invention provides a method of making a set of arrays of fragments of DNA of interest, which method comprises:
a) selecting, from a set of n restriction endonuclease enzymes, a ~s subset of r restriction endonuclease enzymes;
b) digesting genomic DNA with the subset of r enzymes;
c) ligating to the resulting fragments restriction-enzyme-cutting-site-specific adapters with unique polymerase chain reaction amplifiable sequences;
?o d) splitting the resulting fragments into r2 aliquots;
e) amplifying each aliquot with two-restriction enzyme-specific WO 00/55364 PCT/GB00/00916 =

pnmers;
f) forming an array of the r2 aliquots of the amplimers;
and g) repeating steps a) to f) using a different subset of r restriction endonuclease enzymes. The invention also includes sets of arrays obtained or obtainable by the method.
The n restriction endonuclease enzymes may be selected from 4-cutters and 5-cutters and 6-cutters, and a set may include enzymes from one or two or three of these categories. The value of n is preferably 3 ~o to 10, for reasons discussed below. The value of r is less than n and is preferably 2 to 4, chosen with reference to the frequency with which the chosen enzymes cut nucleic acids, and ease of fragment amplification by PCR.
Is Terminal restriction site profiling arrays (TRSPAs) If we use a total of n 6 by cutter restriction enzymes within the total set of enzymes used for fragmentation, let us use subsets of r 6 by cutter enzymes (taken from the total set n) to make (rxr) TRSPA test matrices as follows.
~o Version 1 - (TRSPA-1 ) For each (rxr) TRSPA test matrix, digest DNA to completion with all r restriction enzymes. Ligate on restriction enzyme cutting-site-specific adaptors with unique polymerase chain reaction amplifiable tags.
~s Split into rz aliquots and for each aliquot, amplify with biotinylated restriction enzymes and non-biotinylated restriction enzymek tag-specific primers and array the non-biotinylated strands for all values of j and k between 1 and r.

Example Consider the following dsDNA
__~p,~__1 __~g~__2__~g~__3__~C~__4__~,q~__5__~g~__ _ where the restriction enzyme cutting sites are denoted A, B
and C and the fragments after restriction digestion are denoted 1, 2, 3, 4 and 5. + and - denote the sense of the strands.
~o Cut to completion with A, B and C to give A--1--B B--2--B B--3--C C--4--A A--5--B +

Polymerase chain reaction amplify according to the following primer matrix and streptavidin capture to give A B C

bio-A bio-A--1--BC--4--A
+ +

A--1--B C--4--A-bio - -bio-A--5--B
+

-bio-B A--1--B bio-B--2--Bbio-B--3--C
~ + +

A--1--B-bioB--2--B B--3--C
- - -+ +

A--5--B-bioB--2--B-bio - -bio-C bio-C--4--AB--3--C
+ +

C--4--A B--3--C-bio - -Keep only the non-biotinylated strands to give A B C

bio-A A--1--BC--4--A
- +

-bio-B A--1--BB--2--BB--3--C
+ - -+ +

bio-C C--4--AB--3--C
- +

In simplified form A B C

bio-A 1-.5- 4+

bio-B 1 2-. 3-+. 5+ 2+

bio-C 4- 3+

Repeat for all ~C~ combinations of restriction enzymes to generate the TRSPA.
TRSPA-1 hybridisation patterns The two types of TRSPA-1 hybridisation pattern we should expect using a probe resulting from inter-population perfectly matched ~o duplex depletion are 1. Hybridisation to an off-diagonal element (e.g. row x, column y - where x and y are different) and its complementary element reflected across the diagonal (i.e. row y, column x) and 2. Hybridisation to an on-diagonal element (e.g. the element at row z, column z).
TRSPA-1 analysis - a worked example for n=3 and r=2 Let us take the dsDNA
?o -C-1-B-2-B-3-A-4-A-5-C-6-C-7-B-8-A-9-A-where A, B and C denote restriction enzyme cutting sites. 1, 2, 3, 4, 5, 6, 7, 8 and 9 denote the restriction fragments after digestion.
'_' S
The TRSPA-1 test matrices There are 3C2=3 TRSPA-1 test matrices - AB, BC and AC.

The test matrix hybridisation patterns There are {(2~(2+1 ))/2}=3 possible hybridisation patterns for each of the test matrices s The combinatorial diversity If there are three possible TRSPA-1 test matrix hybridisation patterns and three possible test matrices, then there will be 33=27 possible TRSPA-1 signatures.
Fragment 4 analysis - the AB TRSPA-1 matrix Take the dsDNA

~s Cut with A and B to give 2o B-3-A

?s The fragment complementary to fragment 4 will be Polymerase chain reaction amplify and hybridise with just fragment 4 to give AB matrix A B
bio-A
bio-B
Fragment 4 analysis - the BC TRSPA-1 matrix Take the dsDNA

io Cut with B and C to give ~ s C-6-C

B-8_A_9_A_ The fragment complementary to fragment 4 will be ?o Polymerase chain reaction amplify and hybridise with just fragment 4 to give ?s BC matrix B C
bio-B
bio-C I
Fragment 4 analysis - the AC TRSPA-1 matrix Take the dsDNA

Cut with A and C to give ~ o -C

A-The fragment complementary to fragment 4 will be Polymerase chain reaction amplify and hybridise with just fragment 4 to give AC matrix A C
bio-A
bio-C

Fragment 5 analysis - the AB TRSPA-1 matrix Take the dsDNA

Cut with A and B to give -C_1-B

~o B-3-A

~s The fragment complementary to fragment 5 will be ~o Polymerase chain reaction amplify and hybridise with just fragment 5 to give AB matrix A B
bio-A
bio-B
l I
Fragment 5 analysis - the BC TRSPA-1 matrix 2s Take the dsDNA

Cut with B and C to give a B-3-A-4-A-5-C

B_8_A-9_A_ to The fragment complementary to fragment 5 will be Polymerase chain reaction amplify and hybridise with just >s fragment 5 to give BC matrix B C
bio-B
bio-C
J~~il' I
Fragment 5 analysis - the AC TRSPA-1 matrix Take the dsDNA
~o Cut with A and C to give ~5 -C

A-The fragment complementary to fragment 5 will be io Polymerase chain reaction amplify and hybridise with just fragment 5 to give AC matrix A C
bio-A
bio-C
Fragment 6 analysis - the AB TRSPA-1 matrix is Take the dsDNA

Cut with A and B to give The fragment complementary to fragment 6 will be Polymerase chain reaction amplify and hybridise with just fragment 6 to give AB matrix A B
bio-A
bio-B
I ~ I
Fragment 6 analysis - the BC TRSPA-1 matrix ~c~ Take the dsDNA

Cut with B and C to give ~o C-7-B
B_8-A_9_A-The fragment complementary to fragment 6 will be ~s C-6-C
Polymerase chain reaction amplify and hybridise with just fragment 6 to give BC matrix B C
bio-B
bio-C
I I 1.
Fragment 6 analysis - the AC TRSPA-1 matrix Take the dsDNA

Cut with A and C to give ~ o -C

~s C-7-B-8-A

A-The fragment complementary to fragment 6 will be Polymerase chain reaction amplify and hybridise with just fragment 6 to give AC matrix A C
bio-A
bio-C
I I

Overall results Fragment 4 TRSPA-1 analysis AB BC AC
s Fragment 5 TRSPA-1 analysis AB BC AC
Fragment 6 TRSPA-1 analysis ~o AB BC AC
Version 2 - (TRSPA-2) For each (rxr) TRSPA test matrix, digest DNA to completion with all r restriction enzymes. Ligate on restriction enzyme cutting-site-~s specific adaptors with unique polymerase chain reaction amplifiable tags.
Split into r2 aliquots and for each aliquot, amplify with non-biotinylated restriction enzymes and non-biotinylated restriction enzymek tag-specific primers and array the denatured strands for all values of j and k between 1 and r.
~o Example Consider the following dsDNA
__~A~__1 __~g~__2__~g~__3__~C~__4__~A~__5__~g~_- +
__~A~__1__~g~__2__~g~__3__~C~-_4__~A~__5__~g~__ _ where the restriction enzyme cutting sites are denoted A, B
and C and the fragments after restriction digestion are denoted 1, 2, 3, 4 and 5. + and - denote the sense of the strands.
~o Cut to completion with A, B and C to give A--1--B B--2--B B--3--C C--4--A A--5--B +

Is Polymerase chain reaction amplify according to the following primer matrix to give A B C

+ +

- -+

g__2__g _ +

-+ + +

A__1__gg__2__g g__2__g _ _ _ + +

g__2__g g__3__C
_ _ +

-+ +

- -+

-In simplified form A B C
A 1,2,5 4 B 1,2,5 2 2,3 C 4 2, 3 Repeat for all "C~ combinations of restriction enzymes to generate the TRSPA.

TSPSA-2 hybridisation patterns The two types of TRSPA-2 hybridisation pattern we should expect using a probe resulting from inter-population perfectly matched duplex depletion are 1. Hybridisation to an off-diagonal element (e.g. row x, column y - where x and y are different) and its complementary element reflected across the diagonal (i.e. row y, column x) and 2. Hybridisation to a whole row and column intersecting at an on-diagonal element (e.g. all of row z and all of column z).
TRSPA-2 analysis - a worked example for n=3 and r=2 Let us take the dsDNA

where A, B and C denote restriction enzyme cutting sites. 1, 2, 3, 4, 5, 6, 7, 8 and 9 denote the restriction fragments after digestion.
2o The TRSPA-2 test matrices There are 3C2=3 TRSPA-2 test matrices - AB, BC and AC.
The test matrix hybridisation patterns There are {(2-(2+1 ))/2}=3 possible hybridisation patterns for ~s each of the test matrices The combinatorial diversity If there are three possible TRSPA-2 test matrix hybridisation patterns and three possible test matrices, then there will be 33=27 possible TRSPA-2 signatures.
Fragment 4 analysis - the AB TRSPA-2 matrix s Take the dsDNA

Cut with A and B to give io ~ s A-5-C-6-C-7-B

The fragment complementary to fragment 4 will be Polymerase chain reaction amplify and hybridise with just fragment 4 to give AB matrix A B
A
B
i Fragment 4 analysis - the BC TRSPA-2 matrix Take the dsDNA

Cut with B and C to give B_8_A_9_A_ ~o The fragment complementary to fragment 4 will be ~s Polymerase chain reaction amplify and hybridise with just fragment 4 to give BC matrix B C
B
C
Fragment 4 analysis - the AC TRSPA-2 matrix zo Take the dsDNA

Cut with A and C to give -C

a A-The fragment complementary to fragment 4 will be ~o Polymerase chain reaction amplify and hybridise with just fragment 4 to give AC matrix A C
A
C
I I
~s Fragment 5 analysis - the AB TRSPA-2 matrix Take the dsDNA

2o Cut with A and B to give ?s A-4-A

The fragment complementary to fragment 5 will be Polymerase chain reaction amplify and hybridise with just fragment 5 to give AB matrix A B
A
B
Fragment 5 analysis - the BC TRSPA-2 matrix ~o Take the dsDNA

Cut with B and C to give ~s ?o C-7-B
B_8_A_9_A_ The fragment complementary to fragment 5 will be ~s B-3-A-4-A-5-C
Polymerase chain reaction amplify and hybridise with just fragment 5 to give BC matrix B C
B
C
Fragment 5 analysis - the AC TRSPA-2 matrix Take the dsDNA
s Cut with A and C to give ~ o -C

~ s C-7-B-8-A

A-The fragment complementary to fragment 5 will be Polymerase chain reaction amplify and hybridise with just fragment 5 to give AC matrix A C
A
C

WO 00/55364 PCT/GB00/00916 =

Fragment 6 analysis - the AB TRSPA-2 matrix Take the dsDNA

J
Cut with A and B to give B-2_B
~o B-3-A

The fragment complementary to fragment 6 will be 2o Polymerase chain reaction amplify and hybridise with just fragment 6 to give AB matrix A B
A
B
Fragment 6 analysis - the BC TRSPA-2 matrix ?s Take the dsDNA

Cut with B and C to give s B-3-A-4-A-5-C

B_8_A_9_A_ o The fragment complementary to fragment 6 will be Polymerase chain reaction amplify and hybridise with just is fragment 6 to give BC matrix B C
B
C
l Fragment 6 analysis - the AC TRSPA-2 matrix Take the dsDNA

Cut with A and C to give 's -C

WO 00/55364 PCT/GB00/00916 =

A-s The fragment complementary to fragment 6 will be ~o Polymerase chain reaction amplify and hybridise with just fragment 6 to give AC matrix A C
A
C
I
Overall results is Fragment 4 TRSPA-2 analysis AB BC AC
Fragment 5 TRSPA-2 analysis AB BC AC
i ~ i Fragment 6 TRSPA-2 analysis AB BC AC
The number of test matrices For a total of n enzymes used for fragmentation and a panel of r enzymes per (rxr) test matrix, there will be ~C~ possible (rxr) test matrices, where nC~ - (nl)/[(n-r)l.rl~
~o nCr for various n and r is given in the following table n=

r 3 4 5 6 7 8 9 10 10- _ _ - _ _ _ 1 The number of TRSPA spots arrayed is The total number of TRSPA spots is given by ~-{nCr~ = r 2v(n!)~((n-r)!~r!J~

WO 00/55364 PCT/GB00/00916 =

r2~{nC~} for various n and r is given in the following table n=

1 ( 3 4 5 6 7 8 9 10 3 9 36 90 180315504 756 1,080 ~

4 - 16 80 2405601,1202.0163.360 - - 25 1505251,4003,1506,300 6 - - - 36 2521,0083.0247.560 7 - - - - 49 392 1.7645.880 8 - - - - -~ 64 576 2.880 I

g _ _ _ _ _ _ g1 810 ~

Each test as above will give rise to a particular hybridisation s 'signature'.

(for n=6 and r=3) (for n=6 and r=3) There will be {1+2+...+r} _ {r(r+1)/2} patterns per (rxr) test matrix. If there are nC~ (rxr) test matrices, the total possible number of signatures will be given by {r(r+1 )/2} raised to the power "Cr. We require ~s that the number of such different signatures is greatly in excess over the number of fragments generated upon fragmentation.
If we choose n=6 and r=3, there will be 6 possible hybridisation patterns per test and 20 possible tests - giving a combinatorial diversity of 6'° = 3.7~10'F. The preferred scheme therefore employs six enzymes, tested in groups of three - giving 180 spots per experiment. In order to avoid restriction fragment length polymorphism problems, a duplicate analysis could be performed with a different set of enzymes sharing no cutting sites in common with the first set.
m Example restriction enzyme sets for the preparation of test matrices for TRSPAs Criteria for enzyme choice In order to make TRSPAs, the selection of suitable enzymes ~s is an important factor. Ideally, two sets of different enzymes are required to eliminate the small possibility that a phenotype-determining polymorphism might fall within a chosen restriction site and therefore compromise the specificity of the resulting signature. The selection of enzymes can be based upon a number of criteria ~c> a) The enzymes should be 6 by cutters.
b) Cleavage by any selected enzyme should leave a 4 by overhang at the 5' end.
c) The selected enzymes in each set should all work efficiently under the same buffer conditions.
~s d) The selected enzymes in each set should ideally work efficiently at a single incubation temperature.
e) The chosen enzymes should be commercially available -ideally at concentrations of 10 U / ~I or more.
f) The 5' overhangs left by any two enzymes in the same set ~o should not be identical.
g) No enzyme should appear in both sets for TRSPA fabrication.

h) Enzymes should be selected to avoid or minimise the effects of mammalian methylation patterns. In particular, enzymes with CG
dinucleotides in their recognition sites should be avoided unless the enzyme is known to be able to restrict mSCpG sites.
DNA methylation In vertebrates, DNA is often methylated at the 5t" position of cytosine in the sequence of CpG and this is the only chemical modification that DNA of vertebrates contains under physiological conditions. By the m careful selection of enzymes which do not contain CpG sequences within the recognition site, or the selection of enzymes which freely restrict m~CpG
methylated sites, it is possible to remove the potentially adverse effects of DNA methylation from the TRSPA analysis. 6 by cutters which are known to restrict m~CpG modified DNA efficiently to leave a 5' four base overhang ~s are BspEl and Xmal. These enzymes are therefore the only enzymes with restriction sites containing CpG dinucleotides that are potentially useful in a TRSPA analysis.
Enzyme selection method ~o Sixteen possible four base pair overhangs exist (excluding unusual enzymes with asymmetrical recognition sequences such as BssSl or Bsin, five of which contain CG in the sequence. A further four overhangs could potentially contain CG sequences within the restriction recognition site if preceded by a C and followed by a G. Enzymes are ~s therefore preferentially selected from the remaining six groups.
Excluding isoschizomers, there are up to four possible enzymes which would leave a particular 5' four by overhang. For example, enzymes leaving a CTAG overhang are ~o Avrll - CCTAGG

Nhel - GCTAGC

Spel - ACTAGT

Xbal - TCTAGA

Enzymes to cleave sites with all the combinations of flanking bases are not available for all overhangs - hence the enzyme choice is more limited for some overhang groups than others.
As a primary step towards enzyme selection. the enzymes i c~ supplied by Amersham Pharmacia Biotech, New England Biolabs and Promega are ordered below by overhang sequence. Supplementary details such as the percentage activity in common buffers, the reaction temperature, concentration and supplier are also recorded. For three overhang sequences, there are no available enzymes, the remaining 13 ~s are described below. Enzymes considered unsuitable due to methylation sensitivity are shaded (darker shade). Enzymes considered unfavourable due to the presence of CpG sites even though they do restrict methylated DNA to some degree are also shaded (lighter shaded).
~o Candidate restriction enzymes leaving a 5' overhang of GTAC
Enzyme Site H/M!T buffer Optimu U Supplier Inactivate + BSA /
~I

activity m C C/min Acc65I GGTACC 10/50/100 37 10 NEB 65 / 20 BsiW! CGTACG 100(50)/100(50)/25{-)55 %..12NE~~
{37) ~

; ~ ~~
~
~,~.

BsrGl TGTACA 10/100/100 37 10 NEB 80 / 20 Spll CGTACG 10012016 : 55: 1 ~ :<:

d J~~ l ... ~ E ,x. S%~ ~ . ~'L y'n.
. ; ~Bs~W~~ -Y : '~' x ~ C ~
'~L ~

' FYw y..
'~ ~C
~

WO 00/55364 PCT/GB00/00916 =

Candidate restriction enzymes leaving a 5' overhang of TTAA
Enzyme Site H/M/T buffer Optimu U / ul SupplierInactivate + BSA

o activity m ~C 'C/min Aflll CTTAAG 25/100/100 37 10 APB 60 /

Candidate restriction enzymes leaving a 5' overhang of AATT
Enzyme Site H/M~T buffer Optimu U / ~I Supplier Inactivate + BSA m C C/min o activity EcoRl GAATTC 100/100/100 37 >40 APB/NEB 65 / 20 Mfel CAATTG 10/50/100 37 10 NEB 65 / 20 Candidate restriction enzymes leaving a 5' overhang of CATG
Enzyme Site H/M/T buffer Optimu U SupplierInactivate + BSA m C / C/min activity ~I

BspHl TCATGA 50/100/100 37 10 NEB/APB 65 /

Ncol CCATGG 100/100/100 37 50 NEB/APB 65 /

~c~ Candidate restriction enzymes leaving a 5' overhang of GATC
Enzyme Site H/M/T buffer Optimu U SupplierInactivate + BSA m C / C/min activity ~I

BamHl GGATCC 50/100/75 37 >40 APB/NEB 80 /

Bcll TGATCA 100/100/75 50 (37)10 NEB No Bglll AGATCT 100/75/10 37 >40 APB/NEB No Candidate restriction enzymes leaving a 5' overhang of CCGG
Enzyme Site H/M/T buffer Optimu U / SupplierInactivate + BSA m 'C ul 'C/min o activity BspEl TCCGGA 100/10/0 37 10 NEB 80 / 20 NgoMlV GCCGGC 10/501100 37 10 NEB 8f~~:i,;

Xmal CCCGGG 0/100/100 37 10 NEB 65 / 20 Candidate restriction enzymes leaving a 5' overhang of GCGC
Enzyme Site H/M!T buffer Optimu U / SupplierInactivate + BSA ul ,r activity m C C/min Kasl GGCGCC 75/10017x 37 5 NEB 6512t?,~' Candidate restriction enzymes leaving a 5' overhang of TCGA
Enzyme Site H/M/T buffer Optimu U / pl Supplier Inactivate + BSA m C C/min ,' activity PaeR71 CTCGAG 10/100/10t#: 37 20. ., .,, ,NEB

Sal! GTCGAC 50/0/0:, 37 >40 APBINEB fi4~
,.i Xhol CTCGAG 100/100/100 37 >40 APB/NEB f~ ~s~~#~

to Candidate restriction enzymes leaving a 5' overhang of AGCT
Enzyme Site H/M/T buffer Optimu U / SupplierInactivate + BSA pl activity m C C/min Hindlll AAGCTT 10/100/50 37 >40 NEB/APB 65 /

Candidate restriction enzymes leaving a 5' overhang of CGCG
Enzyme Site H/M/T buffer Optimu U SupplierInactivate + BSA /
ul r activity m =C C/min BssHll GCGCGC 1001100/100 50 20 NEB/APB $(~,~s2d' Mlul ACGCGT 100/75/50 37 10 APB/NEB 65~
.. .:..,..d Candidate restriction enzymes leaving a 5' overhang of GGCC
Enzyme Site H/M~ buffer Optimu U Supplier Inactivate + BSA ' /
~I

o activity m =C C/min Eagl CGGCCG 100/25/10 37 50 .
NEED ~5 F
KRIt .,x ~G
k.

Eco521 CGGCCG 20/20/20 37 1f3 AP~3.

(Eag!) "' ' Candidate restriction enzymes leaving a 5' overhang of TGCA
Enzyme Site H/M/T buffer + BSA U SupplierInactivate Optimu / C/min activity m C ~I

ApaLl GTGCAC 10/100/100 37 10 NEB/APB 70 / 15 AIw441 GTGCAC -/50/- ' 37 10 Promega 70 / 15 m Candidate restriction enzymes leaving a 5' overhang of CTAG
Enzyme Site H/M/T buffer Optimu U SupplierInactivate + BSA m C l C/min ~o activity pl Avrll CCTAGG 50/100/100 37 4 NEB No Nhel GCTAGC 10/100/100 37 10 APB/NEB 65 / 20 Spel ACTAGT 25/100/75 37 50 NEB 65 / 20 Xbal TCTAGA 75/100/75 37 100 NEB 65 ! 20 Blnl(Avrll)CCTAGG 40/20/20 37 10 APB No From the above tables. a short-list of the most useful enzymes is given below for each of the buffer conditions shown Overhang Buffer Buffer Buffer H + M + BSA T + BSA
BSA Enzyme Enzyme Enzyme Efficiency Efficiency Efficiency AATT EcoRl 100 EcoRl 100 Mfel 100 Mfel 50 - -CATG Ncol 100 Ncol 100 ' Ncol 100 BspHl 50 BspHl 100 i BspHl 100 GATC ' BgIII 100 BamHl 100 ! BamHl 75 i BamHl 50 Bglll 75 - -TATA - - - - ' - -ACGT - - - - - -CCGG BspEl 100 Xmal 100 Xmal 100 GCGC - - - - - -TCGA - - - - - -AGCT - - Hindlll 100 Hindlll 50 CGCG - - - - - -GGCC - - - - - -TGCA - - ApaLl 100 ApaLl 100 ATAT - - - - - -CTAG Xbal 75 Xbal 100 Nhel 100 Avrll 50 Spel 100 Avrll 100 Blnl 40 Nhel 100 Xbal 75 - - Avrll 100 Spel 75 GTAC Acc65l 100 BsrGl 100 BsrGl 100 Acc651 75 - -TTAA - - Aflll 100 Aflll 100 Taking into account all of the criteria explained above, an example selection of two six-enzyme sets is described below. Reserve enzymes. which could also be used, are shown as well. These reserve enzymes can be substituted (provided this does not cause overhang duplication) if practical problems regarding enzyme availability or performance should occur.
Example enzyme set 't - for restriction in buffer M + BSA
Number Enzyme Site M+BSA Optimum U Supplier Inactivate oho =C / C/min activity ~I

1 BamHl GGATCC 100 37 100 NEB 80 /

2 BsrGl TGTACA 100 37 10 NEB 80 /

3 HindlllAAGCTT 100 37 >40 APB 65 /

4 Ncol CCATGG 100 37 50 NEB 65 /

Spel ACTAGT 100 37 50 NEB 65 /

6 Aflll CTTAAG 100 37 10 APB 60 /

io Example enzyme set 2 - for restriction in buffer M + BSA
Number Enzyme Site M+BSA Optimum U Supplier Inactivate =C / C/min activity ul 1 EcoRl GAATTC 100 37 >40 APB 65 /

2 BspHl TCATGA 100 37 10 APB 65 /

3 Bglll AGATCT 75 37 >40 APB No 4 Xbal TCTAGA 100 37 100 NEB 65 /

5 Acc65l GGTACC 75 37 10 NEB 65 /

6 ApaLl GTGCAC 100 37 10 NEB No 1. Example reserve enzymes Number Enzyme Site M+BSA Optimum U Supplier Inactivate =C / C/min activity ul r1 Xmal CCCGGG 100 37 10 NEB 65 /

r2 BspEl TCCGGA 100 37 10 NEB 80 /

r3 Nhel GCTAGC 100 37 10 APB 65/20 r4 Avrll CCTAGG 100 37 4 NEB No Example specific (3x3) terminal restriction site profiling array (TRSPA) test matrices for above two sets of six enzymes Triplet combinations from BamHl, BsrGl, Hindlll, Ncol, Spel and Aflll For example set 1 - BamHl, BsrGl, Hindlll, Ncol, Spel and Aflll - the 20 (= 6C3) triplet combinations are ~o BamHl BsrGl BamHl Hindlll BsrGl Hindlll BsrGl Spel Hindlll Spel Ncol Aflll BamHl BsrGl BamHl Hindlll BsrGl Hindlll Hindlll Ncol Ncol Aflll Spel Spel BamHl BsrGl BamHl Ncol SpelBsrGl Hindlll Hindlll Ncol Spel Aflll Aflll BamHl BsrGl BamHl Ncol AflllBsrGl Ncol Hindlll Spel Aflll Spel Aflll BamHl Hindlll BamHl Spel AflllBsrGl Ncol Ncol Spel Ncol. Aflll Aflll Triplet combinations from EcoRl, BspHl, Bglll, Xbal, Acc651 and ApaLl For example set 2 - EcoRl, BspHl, Bglll, Xbal, Acc651 and ApaLl - the 20 (= 6C3) triplet combinations are ~s EcoRl BspHl EcoRl Bglll BspHl Bglll BspHl Acc651 Bglll Acc65l Xbal ApaLl EcoRl BspHl EcoR! Bglll BspHl Bglll Bglll Xbal Xbal Apall Acc65l Acc65l EcoRl BspHl EcoRl Xbal Acc651BspHl Bglll Bglll Xbal Acc651 ApaLI ApaLl EcoRl BspHl EcoRl Xbal ApaLlBspHl Xbal Acc651Bglll Acc65l ApaLl ApaLl EcoRl Bglll EcoRl Acc651 BspHl Xbal ApaLlXbal Acc651 Xbal ApaLl ApaLl Example TRSPA-1 test matrices for set 1 - BamHl, BsrGl, Hindlll, Ncol, Spel and Aflll matrix 1 5'-HO-BamHl-primer5'-HO-BsrGl-primer5'-HO-HindJll-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-BsrGl-primer4 5 6 I

5'-biotin-Hindlll-primer7 8 9 matrix 2 5'-HO-BamHl-primer5'-HO-BsrGl-primer5'-HO-Ncol-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-BsrGl-primer4 5 6 5'-biotin-Ncol-primer7 8 9 matrix 3 5'-HO-BamHl-primer5'-HO-BsrGl-primer5'-HO-Spel-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-BsrG/-primer4 5 6 5'-biotin-Spel-primer7 8 9 matrix 4 5'-HO-BamHl-primer5'-HO-BsrGl-primer5'-HO-Aflll-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-BsrGl-primer4 5 6 5'-biotin-Aflll-primer7 8 9 matrix 5 5~-HO-BamHl-primer5'-HO-Hind/ll-primer5'-HO-Ncol-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-Hindlll-primer4 5 6 5'-biotin-Ncol-primer7 8 9 matrix 6 5'-HO-BamHl-primer5'-HO-Hindlll-primer5'-HO-Spel-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-Hindlll-primer4 5 6 5'-biotin-Spel-primer7 8 9 matrix 7 5'-HO-BamHl-primer5'-HO-Hindlll-primer5'-HO-Aflll-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-Hindlll-primer4 5 6 5'-biotin-Aflll-primer7 8 9 matrix 8 5'-HO-BamHl-primer5'-HO-Ncoi-primer5'-HO-Spel-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-Ncol-primer4 5 6 5'-biotin-Spel-primer7 8 9 matrix 9 5'-HO-BamHl-primer5 HO-Ncol-primer 5'-HO-Aflll-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-Ncol-primer4 5 6 5'-biotin-Aflll-primer7 8 9 matrix 10 5'-HO-BamHl-primer5'-HO-Spel-primer5'-HO-Aflll-primer 5'-biotin-BamHl-primer1 2 3 5'-biotin-Spel-primer4 5 6 5'-biotin-Aflll-primer7 8 9 matrix 11 5'-HO-BsrGl-primer5'-HO-Hindlll-primer5'-HO-Ncol-primer 5'-biotin-BsrGl-primer1 2 3 5'-biotin-Hindlll-primer4 5 6 5'-biotin-Ncol-primer7 8 9 matrix 12 5'-HO-BsrGl-primer5'-HO-Hindlll-primer5'-HO-Spel-primer 5'-biotin-BsrGl-primer1 2 3 5'-biotin-Hindlll-primer4 5 6 5'-biotin-Spel-primer7 8 9 matrix 13 5'-HO-BsrGl-primer5'-HO-Hindlll-primer5'-HO-Aflll-primer 5'-biotin-BsrGl-primer1 2 3 5'-biotin-Hindlll-primer4 5 6 5'-biotin-Aflll-primer7 8 9 matrix 14 5'-HO-BsrGl-primer5'-HO-Ncol-primer5'-HO-Spel-primer 5'-biotin-BsrGl-primer1 2 3 5'-biotin-Nco!-primer4 5 6 5'-biotin-Spel-primer7 8 9 matrix 15 5'-HO-BsrGl-primer5'-HO-Ncol-primer5'-HO-Aflll-primer 5'-biotin-BsrGl-primer1 2 3 5'-biotin-Ncol-primer4 5 6 5'-biotin-Aflll-primer7 8 9 matrix 16 5'-HO-BsrGl-primer5'-HO-Spel-primer5'-HO-Aflll-primer 5'-biotin-BsrGl-primer1 2 3 5'-biotin-Spel-primer4 5 6 5'-biotin-Aflll-primer7 8 9 matrix 17 5'-HO-Hindlll-primer5'-HO-Ncol-primer5'-HO-Spe/-primer 5'-biotin-Hindlll-primer1 2 3 5'-biotin-Ncol-primer4 5 6 5'-biotin-Spel-primer7 8 9 matrix 18 5'-HO-Hmdlll-primer5'-HO-Ncol-primer~ 5'-HO-Aflll-primer 5'-biotin-Hindlll-primer1 2 3 5'-biotin-Nco/-primer4 5 6 5'-biotin-Aflll-primer7 8 9 matrix 19 5'-HO-Hindlll-primer5'-HO-Spel-primer5'-HO-Aflll-primer 5'-biotin-Hindlll-primer1 2 3 5'-biotin-Spel-primer4 5 6 5'-biotin-Aflll-primer7 8 9 matrix 20 5'-HO-Ncol-primer5'-HO-Spel-primer5'-HO-Aflll-primer 5'-biotin-Ncol-primer1 2 3 5'-biotin-Spel-primer4 5 6 5'-biotin-Aflll-primer7 8 9 Example TRSPA-1 test matrices for set 2 - EcoRl, BspHl, Bglll, Xbal, Acc651 and ApaLi matrix 21 5'-HO-EcoRl-primer 5'-HO-BspHl-primer5'-HO-Bglll-primer 5'-biotin-EcoRl-primer1 2 3 5'-biotin-BspNl-primer4 I 5 6 5'-biotin-Bglll-primer7 8 9 matrix 22 5'-HO-EcoRl-primer5'-HO-BspNl-primer5'-HO-Xbal-primer 5'-biotin-EcoRl-primer1 2 3 5'-biotin-BspHl-primer4 5 6 5'-biotin-Xbal-primer7 8 9 matrix 23 5'-HO-EcoRl-primer5'-HO-BspHl-primer5'-HO-Acc651-primer 5'-biotin-EcoRl-primer1 2 3 I

5'-biotin-BspHl-primer4 5 6 ~
I

5'-biotin-Acc651-primer7 8 9 matrix 24 5'-HO-EcoRl-primer5'-HO-BspHl-primer5'-HO-ApaLl-primer 5'-biotin-EcoRl-primer1 2 3 5'-biotin-BspHl-primer4 5 6 5'-biotin-ApaLl-primer7 8 9 matrix 25 5'-HO-EcoRl-primer5'-HO-Bglll-primer 5'-HO-Xbal-primer 5'-biotin-EcoRl-primer1 2 3 I

5'-biotin-Bglll-primer4 5 6 5'-biotin-Xbal-primer7 8 9 matrix 26 5'-HO-EcoRl-primer5'-HO-Bglll-primer5'-HO-Acc65l-primer 5'-biotin-EcoRl-primer1 2 3 5'-biotin-Bglll-primer4 5 6 5'-biotin-Acc651-primer7 8 9 matrix 27 5'-HO-EcoRl-primer5'-HO-Bglll-primer5'-HO-ApaLl-primer 5'-biotin-EcoRl-primer1 2 3 5'-biotin-Bglll-primer4 5 6 5'-biotin-ApaLl-primer7 8 9 matrix 28 5'-HO-EcoRl-primer5'-HO-Xbal-primer5'-HO-Acc651-primer 5'-biotin-EcoRl-primer1 2 3 5'-biotin-Xbal-primer4 5 6 5'-biotin-Acc65l-primer7 8 9 matrix 29 5'-HO-EcoRl-primer5'-HO-Xbal-primer5'-HO-ApaLl-primer 5'-biotin-EcoR/-primer1 2 3 5'-biotin-Xbal-primer4 5 6 5'-biotin-ApaLl-primer7 8 9 matrix 30 5'-HO-EcoRl-primer5'-HO-Acc651-primer5'-HO-ApaLl-primer 5'-biotin-EcoRl-primer1 2 3 5'-biotin-Acc651-primer4 5 5'-biotin-ApaLl-primer7 8 9 matrix 31 5'-HO-BspNl-primer5'-HO-Bglll-primer5'-HO-Xba!-primer 5'-biotin-BspHl-primer1 2 3 5'-biotin-Bglll-primer4 5 -6 5'-biotin-Xbal-primer7 8 9 matrix 32 5'-HO-BspHl-primer5'-HO-Bglll-primer5'-HO-Acc651-primer 5'-biotin-BspHl-primer1 2 3 5'-biotin-Bglll-primer4 5 6 5'-biotin-Acc65l-primer7 8 9 matrix 33 i 5'-HO-BspHl-primer5'-HO-Bglll-primer5'-HO-ApaLl-primer 5'-biotin-BspHl-primer1 2 3 I

5'-biotin-Bglll-primer4 5 6 5'-biotin-ApaLl-primer7 8 9 matrix 34 5'-HO-BspHl-primer5'-HO-Xbal-primer5'-HO-Acc651-primer 5'-biotin-BspHl-primer1 2 3 5'-biotin-Xbal-primer4 5 6 5'-biotin-Acc651-primer7 8 9 matrix 35 5'-HO-BspHl-primer5'-HO-Xbal-primer5'-HO-ApaLl-primer 5'-biotin-BspHl-primer1 2 3 5'-biotin-Xbal-primer4 5 6 5'-biotin-ApaLl-primer7 8 9 matrix 36 5'-HO-BspHl-primer5'-HO-Acc651-primer5'-HO-ApaLl-primer 5'-biotin-BspNl-primer1 2 3 5'-biotin-Acc65l-primer4 5 6 5'-biotin-ApaLl-primer7 8 9 matrix 37 5'-HO-Bglll-primer5'-HO-Xbal-primer5'-HO-Acc651-primer 5'-biotin-Bglll-primer1 2 3 5'-biotin-Xbal-primer4 5 6 5'-biotin-Acc651-primer7 8 9 matrix 38 5'-HO-Bglll-primer5'-HO-Xbal-primer5'-HO-ApaLl-primer 5'-biotin-Bglll-primer1 2 3 5'-biotin-Xbal-primer4 5 6 5'-biotin-ApaL/-primer7 8 9 matrix 39 5 -HO-Bglll-primer5'-HO-Acc651-primer5'-HO-ApaLi-primer 5'-biotin-Bglll-primer1 2 3 5'-biotin-Acc651-primer4 5 6 5'-biotin-ApaLl-primer7 8 9 matrix 40 5'-HO-Xbal-primer5'-HO-Acc651-primer5'-HO-ApaLl-primer 5'-biotin-Xbal-primer1 2 3 5'-biotin-Acc651-primer4 5 I 6 5'-biotin-ApaLl-primer7 8 9 Example TRSPA-2 test matrices for set 1 - BamHl, BsrGl, Hindlll, Ncol, Spel and Aflll matrix 1 5'-HO-Bami-II-primer5'-HO-BsrGl-primer5'-HO-Hindlll-primer 5'-HO-BamHl-primer1 2 3 5'-HO-BsrGl-primer4 5 6 5'-HO-Hindlll-primer7 8 9 matrix 2 5'-HO-BamHl-primer5'-HO-BsrGl-primer5'-HO-Ncol-primer 5'-HO-BamHl-primer1 2 3 5'-HO-BsrGl-primer4 5 6 5'-HO-Ncol-primer7 8 9 matrix 3 5'-HO-BamHl-primer5'-HO-BsrGl-primer5'-HO-Spel-primer 5'-HO-BamHl-primer1 2 3 5'-HO-BsrGl-primer4 5 6 5'-HO-Spel-primer7 8 9 matrix 4 5'-HO-BamHl-primer5'-HO-BsrGl-primer5'-HO-Aflll-primer 5'-HO-BamHl-primer1 2 3 5'-HO-BsrGl-primer4 5 6 5'-HO-Aflll-primer7 8 9 matrix 5 5'-HO-BamHl-primer5'-HO-Hindlll-primer5'-HO-Ncol-primer 5'-HO-BamHl-primer1 2 3 5'-HO-Hindlll-primer4 5 6 5'-HO-Ncol-primer7 8 9 matrix 6 5'-HO-BamNl-primer5'-HO-Hindlll-primer5'-HO-Spel-primer 5'-HO-BamHl-primer1 I 2 3 5'-HO-Hindlll-primer4 5 6 5'-HO-Spel-primer7 8 9 matrix 7 5'-HO-BamHl-primer5'-HO-HindJll-primer5'-HO-Aflll-primer 5'-HO-BamHl-primer1 2 3 5'-HO-Hindlll-primer4 5 6 5'-HO-Aflll-primer7 8 9 matrix 8 5'-HO-BamHl-primer5'-HO-Ncol-primer5'-HO-Spel-primer 5'-HO-Bami-II-primer1 2 3 5'-HO-Ncol-primer4 5 6 5'-HO-Spel-primer7 8 9 matrix 9 5'-HO-BamHl-primer5'-HO-Ncol-primer5'-HO-Aflll-primer 5'-HO-BamHl-primer1 2 3 5'-HO-Ncol-primer4 5 6 5'-HO-Aflll-primer7 8 9 matrix 10 5'-HO-BamHl-primer5'-HO-Spel-primer5'-HO-Aflll-primer 5'-HO-BamHl-primer1 2 3 5'-HO-Spel-primer4 5 6 5'-HO-Aflll-primer7 8 9 matrix 11 5'-HO-BsrGl-primer5'-HO-Nindlll-primer5'-HO-Ncol-primer 5'-HO-BsrGl-primer1 2 3 5'-HO-Hindlll-primer4 5 6 5'-HO-Ncol-primer7 8 9 matrix 12 5'-HO-BsrGl-primer5'-HO-Nindlll-primer5'-HO-Spel-primer 5'-HO-BsrGl-primer1 2 3 5'-HO-Hindlll-primer4 5 6 5'-HO-Spel-primer7 8 9 matrix 13 5'-HO-BsrGl-primer5'-HO-Hindlll-primer5'-HO-Aflll-primer 5'-HO-BsrGl-primer1 2 3 5'-HO-Hindlll-primer4 5 6 5'-HO-Aflll-primer7 8 9 matrix 14 5'-HO-BsrGl-primer5'-HO-Ncol-primer5'-HO-Spei-primer 5'-HO-BsrGl-primer1 2 3 5'-HO-Ncol-primer4 5 6 5'-HO-Spel-primer7 8 9 matrix 15 5'-HO-BsrGl-primer5'-HO-Ncol-primer5'-HO-Aflll-primer 5'-HO-BsrGl-primer1 2 3 5'-HO-Ncol-primer4 5 6 5'-HO-Aflll-primer7 8 9 matrix 16 5'-HO-BsrGl-primer5'-HO-Spel-primer5'-HO-Aflll-primer 5'-HO-BsrGl-primer1 2 3 5'-HO-Spel-primer-_ 4- _ . 5_..

5'-HO-Aflll-primer7 8 9 matrix 17 5'-HO-Hindlll-primer5'-HO-Ncol-primer5'-HO-Spel-primer 5'-HO-Hindlll-primer1 2 3 5'-HO-Ncol-primer4 5 6 5'-HO-Spel-primer7 8 9 matrix 18 5'-HO-Hindlll-primer5'-HO-Ncol-primer5'-HO-Aflll-primer 5'-HO-Hindlll-primer1 2 3 5'-HO-Ncol-primer4 5 6 I 5'-HO-Aflll-primer7 ~ 8 ~ 9 ~

matrix 19 5'-HO-Hindlll-primer5'-HO-Spel-primer5'-HO-Aflll-primer 5'-HO-Hindlll-primer1 2 3 5'-HO-Spel-primer4 5 6 5'-HO-Aflll-primer7 8 9 matrix 20 5'-HO-Ncol-primer5'-HO-Spel-primer5'-HO-Aflll-primer 5'-HO-Ncol-primer1 2 3 5'-HO-Spel-primer4 5 6 5'-HO-Aflll-primer7 8 9 Example TRSPA-2 test matrices for set 2 - EcoRl, BspHl, Bglll, Xbal, Acc65l and ApaLl matrix 21 5'-HO-EcoRl-primer 5'-HO-BspHl-primer5 -HO-Bgll/-primer 5'-HO-EcoRl-primer 1 2 3 5'-HO-BspHl-primer 4 5 6 5'-HO-Bglll-primer 7 8 9 matrix 22 5'-HO-EcoRl-primer5'-HO-BspHl-primer5'-HO-Xbal-primer 5'-HO-EcoRl-primer1 2 3 5'-HO-BspHl-primer4 5 ~ 6 5'-HO-Xbai-primer7 I 8 ~ 9 matrix 23 5'-HO-EcoRl-primer 5'-HO-BspHl-primer5'-HO-Acc651-primer 5'-HO-EcoRl-primer ~ 1 2 3 5'-HO-BspHl-primer 4 5 6 5'-HO-Acc651-primer 7 I 8 ~ 9 matrix 24 5'-HO-EcoRl-primer5'-HO-BspHl-primer5'-HO-ApaLl-primer 5'-HO-EcoRl-primer1 2 3 5'-HO-BspHl-primer4 5 6 5'-HO-ApaLl-primer7 8 9 matrix 25 5'-HO-EcoRl-pnmer5'-HO-Bglll-primer5'-HO-Xbal-primer 5'-HO-EcoRl-primer1 2 3 5'-HO-Bglll-primer4 5 6 5'-HO-Xbal-primer7 8 9 matrix 26 5'-HO-EcoRl-primer5'-HO-Bglll-primer5'-HO-Acc651-primer 5'-HO-EcoRl-primer1 2 3 5'-HO-Bglll-primer4 5 6 5'-HO-Acc651-primer7 8 9 matrix 27 5'-HO-EcoRl-primer5'-HO-Bg/ll-primer5'-HO-ApaLl-primer 5'-HO-EcoRl-primer1 2 3 5'-HO-Bglll-primer4 5 6 5'-HO-ApaLl-primer7 8 9 matrix 28 5'-HO-EcoRl-primer 5'-HO-Xbal-primer5'-HO-Acc651-primer 5'-HO-EcoRl-primer1 I 2 3 5'-HO-Xbal-primer4 5 6 5'-HO-Acc651-primer7 8 9 matrix 29 5 -HO-EccRl-primer5'-HO-Xbal-primer5'-HO-ApaLl-primer 5'-HO-EcoRl-primer1 2 3 5'-HO-Xbal-primer4 5 6 5'-HO-ApaLl-primer7 8 9 matrix 30 5'-HO-EcoRl-primer5'-HO-Acc651-primer5'-HO-ApaLl-primer 5'-HO-EcoRl-primer1 2 3 5'-HO-Acc65l-primer4 5 6 5'-HO-ApaLl-primer7 8 9 matrix 31 5'-HO-BspHl-primer~ 5'-HO-Bglll-primer5'-HO-Xbal-primer 5'-HO-BspHl-primer1 2 3 5'-HO-Bglll-primer4 5 6 5'-HO-Xbal-primer7 8 9 matrix 32 5'-HO-BspHl-primer5'-HO-Bglll-primer5'-HO-Acc651-primer 5'-HO-BspHl-primer1 2 3 5'-HO-Bglll-primer4 5 6 5'-HO-Acc65l-primer7 I 8 9 matrix 33 5'-HO-BspHl-primer5'-HO-Bglll-primer5'-HO-ApaLl-primer 5'-HO-BspHl-primer1 2 3 5'-HO-Bglll-primer4 5 6 5'-HO-ApaLl-primer7 8 9 matrix 34 5'-HO-BspHl-primer5'-HO-Xbal-primer5'-HO-Acc651-primer 5'-HO-BspHl-primer1 2 3 5'-HO-Xbal-primer4 5 6 5'-HO-Acc651-primer7 8 9 matrix 35 5'-HO-BspHl-primer5'-HO-Xbal-primer5'-HO-ApaLl-primer 5'-HO-BspHl-primer1 2 3 5'-HO-Xbal-primer4 5 6 5'-HO-ApaLl-primer7 8 matrix 36 ~ 5'-HO-BspNl-pnmer5'-HO-Acc651-primer5'-HO-ApaLl-primer 5'-HO-BspHl-primer 1 2 3 5'-HO-Acc65l-primer 4 5 6 5'-HO-ApaLl-primer 7 8 9 matrix 37 5'-HO-8glll-primer 5'-HO-Xbal-primer5'-HO-Acc651-primer 5'-HO-Bglll-primer1 2 3 5'-HO-Xbal-primer4 5 6 5'-HO-Acc651-primer7 8 9 matrix 38 5'-HO-Bglll-primer5'-HO-Xbal-primer5'-HO-ApaLl-primer 5'-HO-Bglll-primer1 2 3 5'-HO-Xbal-primer4 5 6 ~

5'-HO-ApaLl-primer7 8 9 matrix 39 5'-HO-Bglll-primer5'-HO-Acc651-primer5'-HO-ApaLl-primer 5'-HO-Bglll-primer1 2 3 5'-HO-Acc651-primer4 5 6 5'-HO-ApaLl-primer7 8 9 matrix 40 5'-HO-Xbal-primer5'-HO-Acc651-primer5'-HO-ApaLl-primer 5'-HO-Xbal-primer1 2 3 5'-HO-Acc651-primer4 5 6 5'-HO-ApaLl-primer7 8 9 Hybridisation patterns There are six possible hybridisation patterns for a given probe fragment from inter-population perfectly matched duplex depletion enrichment with a (3x3) test matrix in TRSPA-1 analysis.
~o There are also six possible hybridisation patterns for a given probe fragment from inter-population perfectly matched duplex depletion enrichment with a (3x3) test matrix in TRSPA-2 analysis.
We can denote these patterns as follows ~s TRSPA-1 analysis Pattern a (1 only) matrix # i 5'-HO-X-primer 5'-HO-Y-primer5'-HO-Z-primer 5'-biotin-X-primer ' 2 3 5'-biotin-Y-primer 4 5 6 5'-biotin-Z-primer 7 8 9 Pattern b (2 and 4) matrix # 5'-HO-X-primer 5'-HO-Y-primer5'-HO-Z-primer 5'-biotin-X-primer1 t~~ I 3 5'-biotin-Y-primer5 6 5'-biotin-Z-primer7 8 9 Pattern c (3 and 7) ~o matrix # 5'-HO-X-primer5'-HO-Y-primer 5'-HO-Z-primer 5'-biotin-X-primer1 2 5'-biotin-Y-primer4 5 6 5'-biotin-Z-primer 8 9 I I I

Pattern d (5 only) matrix # 5'-HO-X-primer 5~-HO-Y-primer5'-HO-Z-primer 5'-biotin-X-primer1 2 3 5'-biotin-Y-primer4 6 5'-biotin-Z-primer7 8 9 IS

Pattern a (6 and 8) matrix # 5'-HO-X-primer 5'-HO-Y-primer5'-HO-Z-primer 5'-biotin-X-primer1 2 3 5'-biotin-Y-primer4 5 5'-biotin-Z-primer7 ~ 9 Pattern f (9 only) matrix # 5'-HO-X-primer ~'-HO-Y-primer 5'-HO-Z-primer 5'-biotin-X-primer1 2 3 I

5'-biotin-Y-primer4 i 5 6 5'-biotin-Z-primer7 8 TRSPA-2 analysis Pattern a (row 1 and column 1 ) matrix # 5'-HO-X-primer ~ 5'-HO-Y-primer 5'-HO-Z-primer 5'-HO-X-primer 5'-HO-Y-primer 5 6 .

.
5'-HO-Z-primer 8 9 Pattern b (2 and 4) matrix # 5'-HO-X-primer 5'-HO-Y-primer 5'-HO-Z-primer 5'- HO -X-primer 1 3 5'- HO -Y-primer 5 6 5'- HO -Z-primer 7 8 9 Pattern c (3 and 7) matrix ~ 5'-HO-X-primer 5'-HO-Y-primer 5'-HO-Z-primer 5'- HO -X-primer 1 2 5'- HO -Y-primer 4 5 6 5'- HO -Z-primer 8 9 I I i Pattern d (row 2 and column 2) s Pattern a (6 and 8) matrix # 5'-HO-X-primer 5'-HO-Y-primer 5'-HO-Z-primer 5'- HO -X-primer1 2 3 5'- HO -Y-primer4 5 5'- HO -Z-primer7 9 iu Pattern f (row 3 and column 3) Signatures If we have 40 such (3x3) test matrices and we denote the ~ s possible hybridisation patterns a, b, c, d, a and f, then we can write the overall TRSPA hybridisation signature as follows WO 00/55364 PCT/GB00/00916 =

Matrix Pattern matrixpattern 1 I a. b, c. d.eorf21 a, b, c, d.eorf 2 a, b, c. d, 22 a, b, c.
a or f d, a or f 3 a. b. c,d,eorf 23 a, b, c,d,eorf 4 a. b, c, d. 24 a, b, c, a or f d, a or f a, b, c, d, 25 a, b, c, a or f d, a or f 6 a, b, c, d, 26 a, b, c, a or f d, a or f 7 a, b, c, d. 27 a, b, c, a or f d, a or f 8 a, b, c, d, 28 a, b, c, a or f I d, a or f 9 a, b, c, d, 29 a, b, c, a or f d, a or f ~

i a. b, c, 30 a, b, c, d. a or f d. a or f 11 a, b, c, d, 31 a, b, c, a or f d, a or f 12 a, b, c,d,aorf32 a, b, c, d,aorf 13 a, b, c, d, 33 a, b, c, a or f d, a or f 14 a, b, c, d, 34 a, b, c, a or f d, a or f a, b, c. d, 35 a, b, c, a or f d, a or f 16 a, b, c, d, 36 a, b, c, a or f d, a or f 17 a, b, c, d, 37 a, b, c, a or f d, a or f 18 a, b, c, d, 38 a, b, c, a or f d, a or f 19 a, b, c, d, 39 a, b, c, a or f d, a or f a, b, c, d, 40 a, b, c, a or f d, a or f Aspect (III) In another aspect the invention provides a nucleic acid characterisation method which comprises presenting to the set of arrays as defined above a nucleic acid fragment of interest under hybridisation conditions, and observing a pattern of hybridisation. Preferably, a plurality of nucleic acid fragments of interest are separately presented to the set of arrays, and the resulting patterns of hybridisation are compared.
m Preferably, the plurality of nucleic acid fragments of interest are drawn from the mixture of DNA fragments, enriched in fragments that are characteristic of a phenotype of interest, as described under the invention (1 ) above.
Thus in this aspect the invention provides a method of identifying fragments of DNA that are characteristic of a phenotype of _77_ interest, which method comprises recovering, cloning and amplifying individual DNA fragments from the mixture of DNA fragments obtained under invention (1 ) above, presenting the individual DNA fragments to the set of arrays as defined under hybridisation conditions, observing a pattern of hybridisation generated by each individual DNA fragment, and subjecting to further investigation any two or more individual DNA fragments whose hybridisation patterns are similar or identical.
TRSPA signatures and whole genome association studies m After a given number of cycles of inter-population perfectly matched duplex depletion, phenotype determining fragments will be enriched but will not be entirely free from 'noise' fragments. Noise may result from unequal allelic frequencies for certain SNPs between the two populations. Noise will also result from the presence of somatic mutations ~s in the cells used to prepare DNA fragments and from the use of polymerase chain reaction in some of the embodiments of the current invention.
For the preferred embodiment, DNA is prepared from a library of clones (either genomic clones or cDNA clones) - with inserts derived =a from the individuals) and propagated in some appropriate host and cloning vector system. Restriction enzyme fragmentation is used prior to cloning and polymerase chain reaction amplification is used to prepare the DNA for comparison in fragmented form. Priming sites within the vector sequence flanking the cloned restriction enzyme fragmented inserts are employed for 2s one or more cycles of polymerase chain reaction amplification of the fragmented DNA of interest. The primers used for polymerase chain reaction amplification of the fragmented DNA of interest are again used after the phenotype-determining fragment enrichment process to 'rescue' and clone the enriched fragments. Cloned enriched fragments are colony ~o purified, picked into appropriate storage containers, catalogued and archived, DNA probes are prepared from these single clones and are _78_ individually hybridised against 180 spot TRSPA arrays as above. TRSPA
signatures are determined for many colonies. Most noise fragments will be random in nature and will thus be randomly distributed amongst the 62° -3.7~10'J possible types of TRSPA signature. Phenotype-determining signal fragments, however, will be those where repeat TRSPA signatures are obtained as more and more colonies are sampled. The more frequently the repetition of a given TRSPA signature occurs per unit colonies sampled, the greater the signal to noise ratio and the more successful has been the en richment.
~c~ A great many of the steps in this process are amenable to high throughput automation - enabling very large numbers of single colony TRSPA signatures to be determined with ease and extending the power of the current invention to cases where signal to noise ratios are beyond current approaches.
~s Statistical correlations (associations) can initially be drawn between TRSPA signatures and phenotype. The clones giving rise to a particular TRSPA signature showing a useful association with a phenotype of interest can then be sequenced in order to determine at a DNA
sequence level the associations) with the phenotype of interest. Such ~o associations have future predictive values for the phenotype of interest, knowing the genotype and will be of great value in medicine and pharmacogenetics.
If the genome of interest is wholly or partially sequenced, we can also in silico restrict the DNA with all n enzymes, calculate the ~s expected signature for each fragment and pattern match these expected signatures with the observed signature (taking into account any loss or gain of restriction sites due to polymorphic variation compared to the reference sequence) to immediately identify the fragment of interest within a gene, genomic region, chromosome or whole genome. This latter method will be ,o of great value in those cases where a great many phenotype-determining fragments are obtained and repeat signatures are rare or unobtained. The _79_ clustering of phenotype-determining fragments to adjacent DNA regions thus gives an association between those genomic regions and the phenotype of interest.
Aspect (IV) In yet another aspect the invention provides a double-stranded DNA molecule having the sequence a-A-b-B...X-y-Y-z where A, B...X and Y are unique restriction sites for n different restriction endonuclease enzymes, and a, b...y, z denotes distances in base pairs, ~ c~ characterised in that each fragment. obtainable by cutting the DNA
molecule by means of any one or more up to n of the restriction enzymes, has a different length from every other fragment.
An example totally diagnostic internal control DNA which allows both ~s the extent and exact nature of any example set 1 (or example set 2) 6 by cutter partial digestion to be unambiguously determined for inter-population perfectly matched duplex depletion or TRSPA restriction In both of the above schemes, it is important that limit digestion products are obtained. Monitoring the extent of partial digestion ~o resulting from multi-enzymatic restriction and determining precisely which enzymes have failed to cut is a task of great importance.
If we have up to six enzymes for DNA digestion - let us label these A, B, C, D, E and F. We need to somehow determine that these have all cut to completion during the fragmentation stage for inter-~s population perfectly matched duplex depletion and also for the digestion step prior to adaptor ligation in TRSPA fabrication. If any of the enzymes have failed to cut to completion, we need to know which ones and to what degree in order to effectively rectify the problem.
~o The structure of the internal control DNA
If we construct a double stranded DNA molecule with the following structure a n d---t---A---u---B---v---C---w--- D---x--- E---y---F---z---a n d where the A, B, C. D, E and F denote the sites for restriction enzyme cutting and t, u, v, w, x, y and z denote distances in base pairs.
This internal control DNA is either uniformly pre-labelled and added to the DNA of interest at an appropriate concentration prior to restriction or is Southern blot probed with a complementary sequence not ~c~ found in the DNA of interest after restriction.
All six enzymes can cut in only one way.
One enzyme can fail to cut in 6C1 = 6 ways, these are: A, B, C, D, E or F failing to cut.
Two enzymes can fail to cut in 6C2 = 15 ways, these are: AB, as AC, AD, AE, AF, BC, BD, BE, BF, CD, CE, CF, DE, DF or EF failing to cut.
Three enzymes can fail to cut in 6C3 = 20 ways, these are:
ABC, ABD, ABE, ABF, ACD, ACE, ACF, ADE, ADF, AEF, BCD, BCE, BCF, BDE, BDF, BEF, CDE, CDF, CEF or DEF failing to cut.
Four enzymes can fail to cut in 6C4 = 15 ways, these are:
~o ABCD, ABCE, ABCF, ABDE, ABDF, ABEF, ACDE, ACDF, ACEF, ADEF, BCDE, BCDF, BCEF, BDEF or CDEF failing to cut.
Five enzymes can fail to cut in 6C5 = 6 ways, these are:
ABCDE, ABCDF, ABCEF, ABDEF, ACDEF or BCDEF failing to cut.
All six enzymes can fail to cut in only one way.
2s Each of the above possibilities will generate one or more fragments from the internal control DNA. If each possible fragment has a discernible size from any other, then we can determine exactly which enzymes have cut and which have not from the size distribution of the fragments generated. The task is therefore to design such a DNA
~o molecule.

WO 00/55364 PCT/GB00/00916 =

Example simulations Seven simulations are given below - varying the size of inter-site fragments. Criteria for a successful outcome include the following 1. The inter-fragment spacing should be greater for larger fragments (so as to aid electrophoretic resolution).
2. All possible fragments should be unambiguously resolvable in size from each-other.
3. Size gaps between bands comprising different numbers of inter-site units should be greater than the size gaps between bands io comprising the same number of inter-site units.
4. The size gaps and size spread from largest to smallest fragment should be electrophoretically compatible.
Simulation 1 ~s Inter-site fragment sizes (in bp) End---A A---B B---C C---D D---E E---F F---end Possible digestion products obtained (in bp) Fragment size gap length to the next (in bp) smallest fragment (in bp) one unit fragments end---A 80 F---end 140 10 two unit fragments end---B 170 30 E---end 270 20 three unit fragments end---C 270 0 D---end 390 30 four unit fragments end---D 380 -10 C---end 500 40 five unit fragments end---E 500 0 B---end 600 50 six unit fragments end---F 630 30 B---end 690 60 seven unit fragments end---end 770 80 Spread = 690 by Simulation 2 Inter-site fragment sizes (in bp) end---A A---B B---C C---D D---E E---F F---end Possible digestion products obtained (in bp) fragment size gap length to the next (in bp) smallest fragment (in bp) one unit fragments end---A 90 A---B i 100 10 F---end 150 10 nvo unit fragments end---B 190 40 E---end 290 20 three unit fragments end---C 300 10 D---end 420 30 four unit fragments end---D 420 0 C---end 540 40 tive unit fragments end---E 550 10 B---end 650 50 six unit fragments end---F 690 40 B---end 750 60 seven unit fragments end---end 840 90 Spread = 750 by Simulation 3 Inter-site fragment sizes (in bp) end---A A---B B---C C---D D---E E---F F---end Possible digestion products obtained (in bp) fragment size gap length to the next (in bp) smallest fragment (in bp) one unit fragments t I

end---A 100 F---end I 160 10 two unit fragments end---B 210 50 E---a n d 310 20 three unit fragments end---C 330 20 D---end 450 30 four unit fragments end---D 460 10 C---and 580 40 five unit fragments end---E 600 20 B---end 700 50 six unit fragments end---F 750 50 B---and 810 60 seven unit fragments end---end 910 100 Spread = 810 by _87_ Simulation 4 Inter-site fragment sizes (in bp) end---A A --B B---C C---D D---E E---F F---end Possible digestion products obtained (in bp) _88_ fragment size gap length to the next (in bp) smallest fragment (in bp) one unit fragments end---A 1 10 F---end 170 10 two unit fragments end---B 230 60 E---end 330 20 three unit fragments end---C 360 30 D---end 480 30 four unit fragments end---D 500 20 C---end 620 40 five unit fragments end---E 650 30 B---end 750 50 six unit fragments end---F 810 60 B---end 870 60 seven unit fragments end---end 980 110 Spread = 870 by _89_ Simulation 5 Inter-site fragment sizes (in bp) end---A A---B B---C C---D D---E E---F F---end Possible digestion products obtained (in bp) f ragment size gap length to the next (in bp) smallest fragment (in bp) one unit fragments end---A 120 F---end 180 10 I

I two unit fragments end---B 250 70 E---end 350 20 three unit fragments end---C 390 40 D---a n d 510 30 four unit fragments end---D 540 30 C---end 660 40 five unit fragments end---E 700 40 B---end 800 50 six unit fragments end---F 870 70 B---end 930 60 seven unit fragments end---end 1050 120 Spread = 930 by Simulation 6 Inter-site fragment sizes (in bp) end---A A---B B---C C---D D---E E---F F---end Possible digestion products obtained (in bp) fragment size gap to length the next (in bp) smallest fragment (in bp) one unit fragments end---A 130 F---end 190 10 two unit fragments end---B 270 80 E---end 370 20 three unit fragments end---C 420 50 D---end 540 30 four unit fragments end---D 580 40 C---end 700 40 five unit fragments end---E 750 50 B---end 850 50 six unit fragments end---F 930 80 B---end 990 60 seven unit fragments end---end 1120 130 Spread = 990 by Simulation 7 Inter-site fragment sizes (in bp) end---A A---B B---C C---D D---E E---F F---end Possible digestion products obtained (in bp) fragment size gap to length the next (in bp) smallest fragment (in bp) one unit fragments end---A 140 F---end 200 10 two unit fragments end---B 290 90 E---end 390 20 three unit fragments end---C 450 60 D---end 570 30 four unit fragments end---D 620 50 C---end 740 40 five unit fragments end---E 800 60 B---end 900 50 six unit fragments end---F 990 90 B---end 1050 60 seven unit fragments end---end 1190 140 Spread = 1050 by According to the above criteria for success, simulation 7 above clearly fulfils all of the requirements.
An example totally diagnostic internal control DNA which allows both the extent and exact nature of any example set 1 (or example set 2) 4 by cutter partial digestion to be unambiguously determined for inter-population perfectly matched duplex depletion If we have up to three enzymes for DNA digestion - let us label these A, B and C. We need to somehow determine that these have ~o all cut to completion during the fragmentation stage for inter-population perfectly matched duplex depletion. If any of the enzymes have failed to cut to completion, we need to know which ones and to what degree in order to effectively rectify the problem.
~s The structure of the internal control DNA
If we construct a double stranded DNA molecule with the following structure a n d---t---A---u---B---v---C---w---a n d where the A, B and C denote the sites for restriction enzyme cutting and t, u, v and w denote distances in base pairs.
This internal control DNA is uniformly pre-labelled and added to the DNA of interest at an appropriate concentration prior to restriction or ~s is Southern blot probed with a complementary sequence not found in the DNA of interest after restriction.
All three enzymes can cut in only one way.
One enzyme can fail to cut in 3C, = 3 ways, these are: A, B or C failing to cut.
~o Two enzymes can fail to cut in 3C2 = 3 ways, these are: AB, AC or BC failing to cut.

All three enzymes can fail to cut in only one way.
Each of the above possibilities will generate one or more fragments from the internal control DNA. If each possible fragment has a discernible size from any other (and from any of the fragments in simulation 7 above for the panel of up to 6 enzymes), then we can determine exactly which enzymes have cut and which have not from the size distribution of the fragments generated. The task is therefore to design such a DNA
molecule.
io Example simulations Six simulations are given below - varying the size of inter-site fragments. Criteria for a successful outcome included the following:
1. The inter-fragment spacing should be greater for larger fragments (so as to aid electrophoretic resolution).
~s 2. All possible fragments should be unambiguously resolvable in size from each-other.
3. Size gaps between bands comprising different numbers of inter-site units should ideally be greater than the size gaps between bands comprising the same number of inter-site units.
~0 4. The size gaps and size spread from largest to smallest fragment should be electrophoretically compatible.
5. The largest fragment obtained should ideally be smaller than the smallest fragment obtained in simulation 7 above for the panel of up to six enzymes.
~5 _97_ Simulation 1 Inter-site fragment sizes (in bp) end---A A---B B---C C---end Possible digestion products obtained (in bp) fragment size gap length to the next (in bp) smallest fragment (in bp) one unit fragments end---A 10 C---end 40 10 two unit fragments end---B 30 -10 B---end 70 20 three unit fragments end---C 60 -10 A---end 90 30 four unit fragments end---end 100 10 Spread = 90 by to Simulation 2 Inter-site fragment sizes (in bp) end---A A---B B---C C---end IS

_98_ Possible digestion products obtained (in bp) fragment size gap to length the next (in bp) smallest fragment (in bp) one unit fragments end---A 15 C---end 45 10 two unit fragments end---B 40 -5 B---end 80 20 three unit fragments end---C 75 -5 A---end 105 30 four unit fragments end---end 120 15 Spread = 105 by Simulation 3 Inter-site fragment sizes (in bp) end---A A---B B---C C---end _99_ Possible digestion products obtained (in bp) fragment size gap length to the next (in bp) smallest fragment (in bp) one unit fragments end---A 20 C---end 50 10 two unit fragments end---B 50 0 B---end 90 20 three unit fragments end---C 90 0 A---and 120 30 four unit fragments end---end 140 20 Spread = 120 by Simulation 4 Inter-site fragment sizes (in bp) end---A A---B B---C C---end Possible digestion products obtained (in bp) fragment size gap length to the next (in bp) smallest fragment (in bp) one umt fragments end---.A 25 C---end 55 I 10 two unit fragments end---B 60 I 5 B---end 100 20 three unit fragments ~

end---C 105 5 A---a n d 135 30 four unit fragments end---end 160 25 Spread = 135 by Simulation 5 Inter-site fragment sizes (in bp) end---A A---B B---C C---end Possible digestion products obtained (in bp) fragment size gap to length the next (in bp) smallest fragment (in bp) one unit fragments end---A 20 C---end 35 5 two unit fragments end---B 45 10 B---end 65 10 three unit fragments end---C 75 10 A---end 90 15 four unit fragments end---end 110 20 Spread = 90 by Simulation 6 Inter-site fragment sizes (in bp) end---A A---B B---C C---end Possible digestion products obtained (in bp) fragment size gap i length to the next yin bp) smallest fragment (in bp) one unit fragments end---A 25 A___g 30 5 C---end I 40 5 two unit fragments end---B 55 I 15 B---end 75 10 three unit fragments end---C 90 15 A---end 105 15 four unit fragments end---end 130 25 Spread = 105 by s According to the above criteria for success, simulation 6 clearly fulfils all of the requirements.
Example determination of the entire set of internal control DNA limit ~o and partial digestion patterns for a panel of up to six restriction enzymes For the example simulation 7, the entire set of internal control DNA limit and partial digestion patterns for a panel of up to six restriction enzymes can be determined as below.
Is Inter-site fragment sizes (in bp) end---A A---B B---C C---D D---E E---F F---end Possible digestion products obtained All six enzymes can cut in only one way.
failed digests by by by by by by by none 140 150 160 170 180 190 200 One enzyme can fait to cut in 6C1 = 6 ways, these are: A, B, C, D, E or F failing to cut.
fo failed by by by by by by digest Two enzymes can fail to cut in 6C2 = 15 ways, these are: AB, AC, AD. AE. AF. BC, BD. BE, BF, CD, CE, CF, DE, DF or EF failing to cut.
failed by by by by by digests Three enzymes can fail to cut in 6C3 = 20 ways, these are:
ABC. ABD. ABE. ABF, ACD, ACE, ACF. ADE, ADF, AEF, BCD. BCE, BCF, BDE, BDF, BEF, CDE, CDF, CEF or DEF failing to cut.
failed by by by digests by J

Four enzymes can fail to cut in 6C4 = 15 ways, these are:
ABCD, ABCE, ABCF, ABDE, ABDF, ABEF, ACDE, ACDF, ACEF, ADEF, BCDE. BCDF, BCEF. BDEF or CDEF failing to cut.
failed by by by digests Five enzymes can fail to cut in 6C5 = 6 ways, these are:
ABCDE, ABCDF, ABCEF, ABDEF, ACDEF or BCDEF failing to cut.
failed by by digests Ail six enzymes can fail to cut in only one way.
failed digests by all 1190 Example determination of the entire set of internal control DNA limit and partial digestion patterns for a panel of up to three restriction ~ o enzymes For the example simulation 6, the entire set of internal control DNA limit and partial digestion patterns for a panel of up to three restriction enzymes can be determined as below.
ua Inter-site fragment sizes (in bp) end---A A---B B---C C---end Possible digestion products obtained All three enzymes can cut in only one way.
failed digests by by by by none 25 30 35 40 One enzyme can fail to cut in 3C, = 3 ways, these are: A. B or C failing to cut.
failed digests by by by Two enzymes can fail to cut in 3C~ = 3 ways, these are: AB, AC or BC failing to cut.
failed digests by by All three enzymes can fail to cut in only one way.
failed digests by all 130 Example 1 a - The digestion internal control plasmid for the 6 by cutter set 1 TRSPA enzymes BamHl, BsrGl, Hindlll, Ncol, Speh and A fill The plasmid pNW33 (shown below) was constructed to contain an insert with all of the 6 by cutter TRSPA enzyme sites.
,y~n~rrr ~4ø~a ~~r ~øss~
140'by~,s~amh'I (605) 150 bp~~ E~srGl 1755) 160 b p X ,yindf II (9'f ~) 170 by ~,r ;v'coi BOSS) 180 by 1 pNUt~3 (3954 bp) '~;N~eI (''1265 190 b p~
(not drai~rn to scale) x,00 by ~t»E ('xø55) ~. Bsp ~1 ('J655) ~ h'rn~rrr ~~sss~
Nindi« j'! 7't5) BspEl sites define the outer ends of the 140 by and the 200 ~o by fragments. The full sequence for pNW33 is shown below:
tcgcgcgtttcggtgatgacggtgaaaacctctgacacatgcagctcccggagacggtcacagcttgtctgtaagcgga tgcc gggagcagacaagcccgtcagggcgcgtcagcgggtgttggcgggtgtcggggctggcttaactatgcggcatcagagc a gattgtactgagagtgcaccatatgcggtgtgaaataccgcacagatgcgtaaggagaaaataccgcatcaggcgccat tc gccattcaggctgcgcaactgttgggaagggcgatcggtgcgggcctcttcgctattacgccagctggcgaaaggggga tgt gctgcaaggcgattaagttgggtaacgccagggttttcccagtcacgacgttgtaaaacgacggccagtgaattcgagc tcgg taccgggccccccctcgaggtcgacggtatcgataagcttgatcgcagctggtaatccggacgcccgcgtcgaagatgt tgg ggtgttgtaacaatatcgattccaattcagcgggggccacctgatatcctttgtatttaattaaagacttcaagcggtc aactatga agaagtgttcgtcttcgtcccagtaaggatccgcactttgaattttgtaatcctgaagggatcgtaaaaacagctcttc ttcaaatct atacattaagacgactcgaaatccacatatcaaatatccgagtgtagtaaacattccaaaaccgtgatggaatggaaca aca cttaaatgtacaccctggtaatccgttttagaatccatgataataattttctggattattggtaattttttttgcacgt tcaaaattttttgca acccctttttggaaacaaacactacggtaggctgcgaaatgttcatactgttgagcaattcacgttcattataagcttt tcactgcat acgacgattctgtgatttgtattcagcccatatcgtttcatagcttctgccaaccgaacggacatttcgaagtattccg cgtacgtg atgttcacctcgatatgtgcatctgtaaaagcaattgttccaggaaccagggcgtatctcttcatagccatggaatacg cctttttc agtgttgcgatgctaatgccgttacaaatattccgagcaccaagaatggctgcgcgcttgcctggtacttgacgtcgta tttgacg gggtccttgagaaagtatttaaactggaacacaatctgaggaatgatcaaagcaaccaacgccaacgcataataactag tg caataccaagacctcccaataatagcacccagacttgtgtaataacctctggctctgatattgctccagatggaattgg acgat atggctcattaattgcgtcgatatctctatcataccagtcgttgattgtctgtgtatagccagtaagacaaggaccaga catcatca tgcaaagaatcgcttaagcccttcttggcctttatgaggatctctctgatttttcttgcgtcgagttttccggtaagac ctttcggtactt cgtccacaaacacaactcctccgcgcaactttttcgcggttgttacttgactggccacgtaatccacgatctctttttc cgtcatcgt ctttccgtgctccaaaacaacaacggcggcgggtccggattaccagctgcgatcaagcttatcgataccgtcgacctcg acct gcaggcatgcaagcttggcgtaatcatggtcatagctgtttcctgtgtgaaattgttatccgctcacaattccacacaa catacga gccggaagcataaagtgtaaagcctggggtgcctaatgagtgagctaactcacattaattgcgttgcgctcactgcccg ctttc cagtcgggaaacctgtcgtgccagctgcattaatgaatcggccaacgcgcggggagaggcggtttgcgtattgggcgct cttc cgcttcctcgctcactgactcgctgcgctcggtcgttcggctgcggcgagcggtatcagctcactcaaaggcggtaata cggtt atccacagaatcaggggataacgcaggaaagaacatgtgagcaaaaggccagcaaaaggccaggaaccgtaaaaag gccgcgttgctggcgtttttccataggctccgcccccctgacgagcatcacaaaaatcgacgctcaagtcagaggtggc gaa I >
acccgacaggactataaagataccaggcgtttccccctggaagctccctcgtgcgctctcctgttccgaccctgccgct taccg gatacctgtccgcctttctcccttcgggaagcgtggcgctttctcatagctcacgctgtaggtatctcagttcggtgta ggtcgttcg ctccaagctgggctgtgtgcacgaaccccccgttcagcccgaccgctgcgccttatccggtaactatcgtcttgagtcc aaccc ggtaagacacgacttatcgccactggcagcagccactggtaacaggattagcagagcgaggtatgtaggcggtgctaca g agttcttgaagtggtggcctaactacggctacactagaaggacagtatttggtatctgcgctctgctgaagccagttac cttcgga ?0 aaaagagttggtagctcttgatccggcaaacaaaccaccgctggtagcggtggtttttttgtttgcaagcagcagatta cgcgc agaaaaaaaggatctcaagaagatcctttgatcttttctacggggtctgacgctcagtggaacgaaaactcacgttaag ggatt ttggtcatgagattatcaaaaaggatcttcacctagatccttttaaattaaaaatgaagttttaaatcaatctaaagta tatatgagt aaacttggtctgacagttaccaatgcttaatcagtgaggcacctatctcagcgatctgtctatttcgttcatccatagt tgcctgact ccccgtcgtgtagataactacgatacgggagggcttaccatctggccccagtgctgcaatgataccgcgagacccacgc tca ?5 ccggctccagatttatcagcaataaaccagccagccggaagggccgagcgcagaagtggtcctgcaactttatccgcct cc atccagtctattaattgttgccgggaagctagagtaagtagttcgccagttaatagtttgcgcaacgttgttgccattg ctacaggc atcgtggtgtcacgctcgtcgtttggtatggcttcattcagctccggttcccaacgatcaaggcgagttacatgatccc ccatgttgt gcaaaaaagcggttagctccttcggtcctccgatcgttgtcagaagtaagttggccgcagtgttatcactcatggttat ggcagc actgcataattctcttactgtcatgccatccgtaagatgcttttctgtgactggtgagtactcaaccaagtcattctga gaatagtgta :~0 tgcggcgaccgagttgctcttgcccggcgtcaatacgggataataccgcgccacatagcagaactttaaaagtgctcat catt ggaaaacgttcttcggggcgaaaactctcaaggatcttaccgctgttgagatccagttcgatgtaacccactcgtgcac ccaac tgatcttcagcatcttttactttcaccagcgtttctgggtgagcaaaaacaggaaggcaaaatgccgcaaaaaagggaa taag ggcgacacggaaatgttgaatactcatactcttcctttttcaatattattgaagcatttatcagggttattgtctcatg agcggataca tatttgaatgtatttagaaaaataaacaaataggggttccgcgcacatttccccgaaaagtgccacctgacgtctaaga aacc ~5 attattatcatgacattaacctataaaaataggcgtatcacgaggccctttcgtc WO 00/55364 PCT/GB00/00916 =

The insert was introduced as a four fragment ligation of 140 (Kpnl-BamHl), 150 / 160 (BamHl-Hindlll), 170 / 180 (Hindlll-Spel) and 190 / 200 (Spel-Xhol) into Kpnl / Sall digested pUCl9c DNA (Genbank X02514). The Sall and Xhol sites were lost as a result of the joining of their compatible sticky ends. The insert can be removed with BspEl or Pvull.
The insert region of 1190 by and the short flanking regions to the vector junctions were sequenced twice in each direction in order to establish the plasmid sequence. In addition, a total of 63 analytical restriction digests and one minus-enzyme control were performed as m detailed in the following table:
Expected band sizes Digest Wellby by by by by by by Uncut 1 1190 BamHl 2 140 1050 BsrGl 3 290 900 Hindlll 4 450 740 Ncol 5 620 570 Spel 6 800 390 Aflll 7 990 200 BamHl BsrGl 8 140 150 900 BamHl Hindlll 9 140 310 740 6amHl Ncol 10 140 480 570 BamHl Spel 11 140 660 390 BamHl Aflll 12 140 850 200 BsrGl Hindlll 13 290 160 740 BsrGl Ncol 14 290 330 570 BsrGISpeI 15 290 510 390 BsrGl Aflll 16 290 700 200 Hindlll Ncol 17 450 170 570 Hindlll Spel 18 450 350 390 Hindlll Aflll 19 450 540 200 Ncol Spel 20 620 180 390 Ncol Aflll 21 620 370 200 Spel AlIII 22 800 190 200 BamHl BsrGl Hindlll 23 140 150 160 740 BamHl BsrGl Ncol 24 140 150 330 570 BamHl BsrGl Spel 25 140 150 510 390 BamHl BsrGl AlIII 26 140 150 700 200 BamHl Hindlll Ncol ~ 140 ~ 170 570 6amHl Hinalll Spei 28 140 310 350 390 BamHl Hindlll Aflll 29 140 310 540 200 BamHl Ncol Spel 30 140 480 180 390 BamHl Ncol Aflll 31 140 480 370 200 BamHl Spel Aflll 32 140 660 190 200 BsrGl Hindlll Ncol 33 290 160 170 570 BsrGl Hindlll Spel 34 290 160 350 390 BsrGl Hrndlll Atlll 35 290 160 540 200 BsrGl Ncol Spel 36 290 330 18C 390 BsrGl NCOI Aflll 37 290 330 370 200 BsrGl Spel Aflll 38 290 510 190 200 Hindlll Ncol Spel 39 450 1 180 390 Hindlll Ncol Aflll 40 450 170 370 200 I I

Nindlll Spel Aflll 41 450 350 190 200 I

Ncol Spel Aflll i 42 620 180 190 200 I

BamHl BsrGl Hindlll 43 140 150 160 170 570 Ncoi ~

BamHl BsrGl Hindlll 44 140 150 160 350 390 Spel BamHl BsrGl Hindlll 45 140 150 160 540 200 Aflll BamHl BsrGl Ncol 46 140 150 330 180 390 Spel BamHl BsrGl Ncol 47 140 150 330 370 200 Aflll BamHl BsrGl Spel 48 140 150 510 190 200 Ailll BamHl Hindlll Ncol 49 140 310 170 180 390 Spel .

BamHl Hindlll Ncol 50 140 310 170 370 200 Aflll BamH'I Hindlll Spel 51 140 310 350 190 200 AlIII

BamHl Ncol Spel Aflll52 140 480 180 190 200 BsrGl Hindlll Ncol 53 290 160 170 180 390 Spel BsrGl Hindlll Ncol 54 290 160 170 370 200 Allll BsrGl Hindlll Spel 55 290 160 350 190 200 Aflll BsrGl Ncol Spel Aflll56 290 330 180 190 200 Hindlll Ncol Spel 57 450 170 180 190 200 Aflll BamHl BsrGl Hindlll 58 140 150 160 170 180 390 Ncol Spel BamHl BsrGl Hindlll 59 140 150 160 170 370 200 Ncol Aflll BamHl BsrGl Hindlll 60 140 150 160 350 190 200 Spel Aflll BamHl BsrGl Ncol 61 140 150 330 180 190 200 Spel Aflll <w ._ BamHl Hindlll Ncol 62 140 310 170 180 190 200 Spel Aflll BsrGl Hindlll Ncol 63 290 160 170 180 190 200 Spel Allll BamHl BsrGl Hindlll 64 140 150 160 170 180 190 200 Ncol Spel Aflll All of these digests produced the expected fragment patterns on agarose gel electrophoresis and a number of these, shaded in the table above, are shown in figure 1. The restriction digests illustrated in figure 1 were carried out using the following conditions:
Digest 1 (Minus enzyme control) pl i 100 ugiml internal control plasmid DNA in 1 x NEB buffer #2 + 100 ug/ml BSA
90 gl ~ 1x NEB buffer #2 + 100 ug/ml BSA
Digest 5 pl ~ 0.33 U/ul Ncol in 1 x NEB buffer #2 + 100 ~g/ml BSA
i i 10 gl i 100 ugiml internal control plasmid DNA in 1 x NEB buffer #2 + 100 gg/ml BSA
75 gl ~ 1 x NEB buffer #2 + 100 gg/ml BSA
Digest 9 >o 15 0.33 U/gl BamHl in 1 x NEB buffer #2 + 100 ~g/ml ~I BSA

15 0.33 U/pl Hindlll in 1 x NEB buffer #2 + 100 gg/ml ~I BSA

10 100 gg/ml internal control plasmid DNA in 1 x NEB
~I buffer #2 + 100 pg/ml BSA

60 1 x NEB buffer #2 + 100 ~g/ml BSA
pl Digest 37 15 0.33 U/~I BsrGl in 1 x NEB buffer #2 + 100 gg/ml gl BSA

15 0.33 U/gl Ncol in 1 x NEB buffer #2 + 100 ~g/ml BSA
gl 15 0.33 U/gl Aflll in 1x NEB buffer #2 + 100 ~g/ml BSA
NI

10 100 gg/ml internal control plasmid DNA in 1 x NEB
~I buffer #2 + 100 gg/ml BSA

45 1 x NEB buffer #2 + 100 ~g/ml BSA
~I

IS

Digest 49 15 I 0.33 Uiul BamHl in 1 x NEB buffer #2 + 100 ug/ml ul BSA

15 I 0.33 U/ul Hindlll in 1 x NEB buffer #2 + 100 gg/ml ul BSA

'; ~ 0.33 U/ul Ncol in 1 x NEB buffer #2 + 100 ug/ml BSA

ul 15 i 0.33 U/ul Spel in 1 x NEB buffer #2 + 100 ug/ml BSA
ul ~ 100 ug/ml internal control plasmid DNA in 1 x NEB
ul buffer #2 + 100 ug/ml BSA

30 ; 1 x NEB buffer #2 + 100 ug/ml BSA
ul Digest 61 i 0.33 U/ul BamHl in 1 x NEB buffer #2 + 100 ~giml ul BSA

i 15 ~ 0.33 U/ul BsrGl in 1 x NEB buffer #2 + 100 gg/ml ul BSA

15 0.33 U/ul Ncol in 1x NEB buffer #2 + 100 gg/ml BSA
pl ~ 15 0.33 U/gl Spel in 1 x NEB buffer #2 + 100 ug/ml BSA
ul i 15 0.33 U/gl Aflll in 1 x NEB buffer #2 + 100 gg/ml BSA
gl 10 100 gg/ml internal control plasmid DNA in 1 x NEB buffer ul #2 + 100 ug/ml BSA

15 1 x NEB buffer #2 + 100 ~g/ml BSA
gl All 64 restriction digests were incubated at 37°C for 6 hours and samples were then electrophoresed on 2.5 % FMC Metaphor agarose gels as illustrated in figure 1. The lanes marked M contain mixed samples ~o from all of the digests 1-64 and these lanes were used as size markers after confirmation of the fragment sizes against Stratagene Kb ladder markers.
When spiked into a genomic DNA digest the internal control restriction fragment pattern produced is indicative of both the degree of t > digestion and the nature of any partial restriction at less than limit digestion.
It is possible to deduce from the bands present which, if any, of the enzymes have failed to cut and therefore to take action to correct this before the DNA is used in a subsequent analysis or enrichment procedure.
~o Preparation of BspE!-released internal control DNA on a large scale The method used for generating internal control PCR product DNA (free from contaminating dNTPs) is depicted in the following figure:
pNW33 D N.::.
crude bio oio + dNTP
~

F
CFc product SA-captured~-SA-bio + dNTP

F~~P ~ uio-SA.-product clean ~-SA-bio F~~~.F:foio-SA~
productf ~
gspEl digest interna~

control for RE
digests Primers and PCR
20 uM B10140UP
l0 5' biotin-CGCAGCTGGTAATCCGGACGCCCGCGTCGAAGATGTT 3' 20 ~M B10200DOWN
5' biotin-CGCAGCTGGTAATCCGGACCCGCCGCCGTTGTTGTT 3' Is Bulk PCR amplification (192x 100 ~I reactions) was carried out according to the conditions described below:
er 20 ml final conditions 100 mM dATP 20 I 100 M

100 mM dCTP 20 ul 100 M

100 mM dGTP 20 I 100 M

100 mM dTTP 20 I 100 M

1 Ox PCR buffer 2 ml 1 x 25 mM MgCl2 400 I 1/50th volume 1 n /ul NW33 400 I 2 ng per PCR

20 uM B10140UP 200 I 20 pmol per PCR

20 uM B10200DOWN 200 ul 20 pmol per PCR

Taq DNA polymerase 100 ul 2.5 U per PCR

The master mix was rapidly dispensed into PCR tubes.
Thermal cycling was initiated using the following parameters: 30 cycles of 97°C for 1 min, 50°C for 2 min, and 72°C for 3 min;
72°C for 5 min; and then 4°C.
After PCR amplification, samples were pooled and subjected to capture of biotinylated PCR product termini and BspEl release of internal control DNA.
m Capture of biotinylated PCR product termini and BspEl release of internal control DNA
All separations were carried out using a Dynal MPC-1 separator (Dynal, product #12001).
20 ml of pooled PCR reaction were mixed with 20 ml of is Dynabeads M-280 in 20 mM tris-HCI (pH 7.4), 2 mM EDTA (pH 8.0), 2 M
NaCI. The tube was incubated at room temperature for 1 hour with rolling.
The Dynabeads were then washed four times with 20 ml of 10 mM tris-HCI (pH 7.4), 1 mM EDTA (pH 8.0), 1 M NaCI.
BspEl digestion was carried out for 1 hour at 37°C in 2 ml of ~c> 0.5 U/yl BspEl in 1 x NEB buffer #3. The digest was then ethanol precipitated by the addition of 0.1 volume of 3 M sodium acetate (pH 5.2) and 2.5 volumes of ethanol, chilling to -20°C and then centrifugation.
Pellets were rinsed with 70 °,'° ethanol prior to redissolution in 500 ~I of 1x TE buffer.
~s 1 pl of the BspEl-released internal control DNA was mixed with 10 ~I of 50 % glycerol AGE loading dye and electrophoresed on a 1.5 agarose gel to confirm that the size of the purified DNA was in accordance with that expected.
~o High molecular weight genomic DNA digestion in the presence of internal control DNA with a dilution series of a mixture of the six set 1 6 by cutters High molecular weight canine genomic DNA was first mixed with a dilution series of a mixture of the six set 1 6 by cutters - each at the same number of units. Aliquots were then removed and mixed with the BspEl-released internal control DNA described above. 20 ug of canine genomic DNA was digested with 0.25, 0.025, 0.0025, 0.00025, 0.000025, 0.0000025 and 0 U/ul BamH1 / BsrGl / Hindlll / Ncol / Spel / Aflll in 200 ~.I.
o> 0.4 yg of canine genomic DNA and 1 ~I of BspEl-released internal control DNA were digested with 0.25, 0.025, 0.0025, 0.00025, 0.000025, 0.0000025 and 0 U/~I BamH1 / BsrGl / Hindlll / Ncol / Spel / Aflll in 4 ~I.
Vistra Green staining was used to monitor the extent of internal control DNA digestion and high molecular weight DNA digestion. The fragment ~s pattern from the internal control DNA is diagnostic of both the degree of digestion and the nature of any partial restriction at less than limit digestion.
A premix of restriction enzymes, buffer and BSA was prepared as detailed below:
component ~I

100 U/~I BamHl 1 ~I

U/~I BsrGl 10 ~I

40 U/~I Hindlll 2.5 ~tl 50 U/~I Ncol 2 ~I

50 U/~I Spel 2 ~I

U/~I Aflll 5 ~I

10x NEB buffer 4 ~I
#2 100x BSA 0.4 ul water 13.
1 ~I

Total volume = 40 ~I
The premix therefore contained each restriction enzyme at 2.5 U/ul in 1 x NEB buffer #2 and 1 x BSA. Serial 10-fold dilutions of this premix were then prepared in 1 x NEB buffer #2 and 1 x BSA.
Premixes of canine genomic DNA, buffer and BSA were also prepared as detailed below:
Component ul 1 mgiml canine genomic 20 ul DNA

10x NEB buffer #2 8 ul 100x BSA 0.8 ul Water ~1.
2 ~1 These premixes therefore contain canine genomic DNA at 0.25 mg/ml in 1 x NEB buffer #2 and 1 x BSA.
Canine genomic DNA and restriction enzyme mixes were ~u then set up as follows:
tube component #1 component #2 80 ~I of canine genomic 20 gl of each restriction DNA at 0.25 enzyme at 2.5 1 mg/ml in 1 x NEB buffer U/pl in 1 x NEB buffer #2 and 1 x BSA #2 and 1 x BSA

80 ul of canine genomic 20 pl of each restriction DNA at 0.25 enzyme at 0.25 2 mg/ml in 1 x NEB buffer U/ul in 1 x NEB buffer #2 and 1 x BSA #2 and 1 x BSA

80 gl of canine genomic 20 ul of each restriction DNA at 0.25 enzyme at 0.025 mg/ml in 1 x NEB buffer U/gl in 1 x NEB buffer #2 and 1 x BSA #2 and 1 x BSA

80 gl of canine genomic 20 ~I of each restriction DNA at 0.25 enzyme at 4 mg/ml in 1 x NEB buffer 0.0025 U/pl in 1 x NEB
#2 and 1 x BSA buffer #2 and 1 x BSA

80 gl of canine genomic 20 gl of each restriction DNA at 0.25 enzyme at mg/ml in 1x NEB buffer #2 0.00025 U/ul in 1x NEB
and ix BSA buffer #2 and 1x BSA

80 gl of canine genomic 20 gl of each restriction DNA at 0.25 enzyme at mg/mt in 1 x NEB buffer 0.000025 U/gl in 1 x NEB
#2 and 1 x BSA buffer #2 and 1 x BSA

80 ul of canine genomic 20 ~I of 1 x NEB buffer DNA at 0.25 #2 and 1 x BSA

mg/ml in 1 x NEB buffer #2 and 1 x BSA

2 yl aliquots were then removed from tubes 1-7 and added to 1 ~~I of BspEl-released internal control DNA and 1 ~I of 2x NEB buffer #2 and 2x BSA to give samples 1-7ic. Samples 1-7ic were then overlaid with 50 ~I of mineral oil in order to prevent evaporation.
The remaining volume from tubes 1-7 was then added to 100 ~I of 1 x NEB buffer #2 and 1 x BSA.
All samples were finally incubated at 37°C overnight.
After digestion, 20 ul of digests 1-7 were mixed with 10 ~I of ici 50 °o glycerol AGE loading dye and 4 ul of digests 1-7ic were mixed with 2 ~I of 50 % glycerol AGE loading dye. Digests in loading dye were then electrophoresed on a 2.5 % Metaphor agarose gel in 1 x TBE. The gel was stained for 60 min in 500 ml of 1 x TBE containing 50 ~I of Vistra Green.
The stained gel was finally imaged on a Fluorimager with the following is settings: a 488 nm laser; a 570 DF 30 filter; a PMT setting of 700 V;
200 ~m resolution; and low sensitivity.
Restriction digests - final conditions Restriction digests 1-7 were therefore performed under the ?o following conditions:
tube~g U U U U U U NEB Nglml canine BamHiBsrGl NindlllNcol Spel Aflllbuffer BSA
genomic #2 DNA

1 20 50 50 50 50 50 50 1x 100 2 20 5 5 5 5 5 5 1 x 100 3 20 0.5 0.5 0.5 0.5 0.5 0.5 1 x 100 4 20 0.05 0.05 0.05 0.05 0.05 0.05 1 x 100 20 0.0050.005 0.005 0.0050.005 0.0051x 100 6 20 0.00050.00050.00050.00050.00050.00051x 100 -- -- ---- -7 20 0 0 0 0 0 ~ ~X ~
0 ,~0 Total volume = 200 ~I

Likewise, restriction digests 1-7ic were performed under the following conditions:
tube ug ul U U U U U U NEB ~g/ml canine IC BamH1 BsrGlHindlll NcolSpel Aflll bufferBSA
aenomic #2 DNA
DNA

tic 0.4 1 1 1 1 1 1 1 1x 100 tic 0.4 1 0.1 0.1 0.1 0.1 0.1 0.1 1 100 x Sic 0.4 1 0.01 0.01 0.01 0.01 0.01 0.01 1 100 x 4ic 0.4 1 0.001 0.0010.001 O.G01 0.0010.001 1x 100 Sic 0.4 ' ~ 0.00010.0001O.OOG1 0.0001C.00010.00011x 100 0.4 0.000010.000010.00001 0.000010.000010.000011x 100 7ic G.4 ~ ~ ~ G ~ 0 0 0 1x 100 IC = BspEl-released internal control DNA
Total volume = 4 ~I
The results are shown in figure 2.

Example 1 b - The digestion internal control plasmid for the 4 by cutter set 1 TRSPA enzymes Haelll, Mbol, and Msel The plasmid pNW35 (shown below) was constructed to contain an insert with all of the 4 by cutter TRSPA enzyme sites.
Hindlll (52) HaeIII [l5) 25 bpi, '~
s0 bp~,. ~~Sel (1051 ~5 by ~ h~lbol (140) L'~N'a'U35 0946 bp) 40 by (riot drawn to scale) ~- E~:uRI (180) Hindlll and EcoRl sites define the outer ends of the 25 by and the 40 by fragments. The sequence of pNW35 is shown below with the inserted region shown in bold type:
atgaccatgattacgccaagctctaatacgactcactatagggaaagcttccggacgtctcaggctaatgttggcccac c gacgttccacgatggggcgctcttaagggcttagaccctcgtcgggagtatttctgtgatctggcgacactcacgcg agaagtcattaccggcgatatgaattcactggccgtcgttttacaacgtcgtgactgggaaaaccctggcgttacccaa ctta atcgccttgcagcacatccccctttcgccagctggcgtaatagcgaagaggcccgcaccgatcgcccttcccaacagtt gcg cagcctgaatggcgaatgggaaattgtaaacgttaatattttgttaaaattcgcgttaaatttttgttaaatcagctca ttttttaacca IS
ataggccgaaatcggcaaaatcccttataaatcaaaagaatagaccgagatagggttgagtgttgttccagtttggaac aag agtccactattaaagaacgtggactccaacgtcaaagggcgaaaaaccgtctatcagggcgatggcccactacgtgaac c atcaccctaatcaagttttttggggtcgaggtgccgtaaagcactaaatcggaaccctaaagggagcccccgatttaga gcttg acggggaaagccggcgaacgtggcgagaaaggaagggaagaaagcgaaaggagcgggcgctagggcgctggcaa gtgtagcggtcacgctgcgcgtaaccaccacacccgccgcgcttaatgcgccgctacagggcgcgtcaggtggcacttt tcg ?0 gggaaatgtgcgcggaacccctatttgtttatttttctaaatacattcaaatatgtatccgctcatgagacaataaccc tgataaat gcttcaataatattgaaaaaggaagagtatgagtattcaacatttccgtgtcgcccttattcccttttttgcggcattt tgccttcctgttt ttgctcacccagaaacgctggtgaaagtaaaagatgctgaagatcagttgggtgcacgagtgggttacatcgaactgga tctc aacagcggtaagatccttgagagttttcgccccgaagaacgttttccaatgatgagcacttttaaagttctgctatgtg gcgcggt attatcccgtattgacgccgggcaagagcaactcggtcgccgcatacactattctcagaatgacttggttgagtactca ccagtc acagaaaagcatcttacggatggcatgacagtaagagaattatgcagtgctgccataaccatgagtgataacactgcgg cc aacttacttctgacaacgatcggaggaccgaaggagctaaccgcttttttgcacaacatgggggatcatgtaactcgcc ttgat cgttgggaaccggagctgaatgaagccataccaaacgacgagcgtgacaccacgatgcctgtagcaatggcaacaacgt t gcgcaaactattaactggcgaactacttactctagcttcccggcaacaattaatagactggatggaggcggataaagtt gcag gaccacttctgcgctcggcccttccggctggctggtttattgctgataaatctggagccggtgagcgtgggtctcgcgg tatcatt gcagcactggggccagatggtaagccctcccgtatcgtagttatctacacgacggggagtcaggcaactatggatgaac ga aatagacagatcgctgagataggtgcctcactgattaagcattggtaactgtcagaccaagtttactcatatatacttt agattgat ttaaaacttcatttttaatttaaaaggatctaggtgaagatcctttttgataatctcatgaccaaaatcccttaacgtg agttttcgttcc I t) actgagcgtcagaccccgtagaaaagatcaaaggatcttcttgagatcctttttttctgcgcgtaatctgctgcttgca aacaaaa aaaccaccgctaccagcggtggtttgtttgccggatcaagagctaccaactctttttccgaaggtaactggcttcagca gagcg cagataccaaatactgtccttctagtgtagccgtagttaggccaccacttcaagaactctgtagcaccgcctacatacc tcgctc tgctaatcctgttaccagtggctgctgccagtggcgataagtcgtgtcttaccgggttggactcaagacgatagttacc ggataa ggcgcagcggtcgggctgaacggggggttcgtgcacacagcccagcttggagcgaacgacctacaccgaactgagatac I ~
ctacagcgtgagctatgagaaagcgccacgcttcccgaagggagaaaggcggacaggtatccggtaagcggcagggtcg gaacaggagagcgcacgagggagcttccagggggaaacgcctggtatctttatagtcctgtcgggtttcgccacctctg actt gagcgtcgatttttgtgatgctcgtcaggggggcggagcctatggaaaaacgccagcaacgcggcctttttacggttcc tggcc ttttgctggccttttgctcacatgttctttcctgcgttatcccctgattctgtggataaccgtattaccgcctttgagt gagctgataccgc tcgccgcagccgaacgaccgagcgcagcgagtcagtgagcgaggaagcggaagagcgcccaatacgcaaaccgcctc tccccgcgcgttggccgattcattaatgcagctggcacgacaggtttcccgactggaaagcgggcagtgagcgcaacgc aat taatgtgagttagctcactcattaggcaccccaggctttacactttatgcttccggctcgtatgttgtgtggaattgtg agcggataa caatttcacacaggaaacagct The 4 by internal control plasmid for restriction enzymes Haelll, Mbol, and Msell was prepared by the insertion of a synthetic 130 by fragment into Hindlll / EcoRl-digested pMOSblue (Amersham Pharmacia Biotech).
The insert region of 130 by was sequenced twice in each direction in order to establish the plasmid sequence. The presence of the ~o restriction sites and the mobility of the fragments released were also checked by restriction digestion and polyacrylamide gel electrophoresis.
Preparation of the internal control spike DNA from the plasmid The following illustration describes the strategy used for the ~s preparation of EcoRl-released internal control DNA:

E:~ Pdb'V'3~
DNA
crude bio + dNTP
PCR product r~A; captured ~-~~-bio + dNTP
PCR product clean ~Sr"~,-bir, PCR product RE digest 4 by cutter internal control for RE digests Primers and PCR
s U-19 mer bio primer 5' bio-GTTTTCCCAGTCACGACGT 3' ICPCR(F) primer 5' TCCGGACGTCTCAGGCTAATGTT 3' Bulk PCR amplification (96x 100 ~I reactions, repeated to ~o give 192 reactions in total) was carried out according to the conditions described below (all volumes are in ~I):
x'1 - (ul) x100 (~I) Final con~l~ti4i W ate r 84.4 8440 -1 Ox PCR buffer 10 1000 1 x 25 mM dNTP mix 0.8 80 200 pM

20 ~M T7 promoter-bio 1 100 20 pmol per primer PCR

20 pM U19 mer-bio primer1 100 20 pmol per PCR

1 ng/pl pNW35 2 I 200 2 ng per PCR

U/~I Taq polymerase 0.8 80 4U per PCR

Total ~- : ~ 100 f ;_ V 1 OOaO . ,.h ~~
.:: y' ., ~~ t ~, ,~ ~i~,=.., f ...~ , .:, _ ., .

The PCR master mix was rapidly dispensed into 96 PCR
tubes. Thermal cycling was initiated using the following parameters:
94°C
for 2 min; 50 'C for 2 min; 29 cycles of 72°C for 2 min, 94°C
for 45 sec, and 50°C for 1 min; 72°C for 8 min; and then 4°C.
Capture of biotinylated PCR product termini and EcoRl- release of internal control DNA
All separations were carried out using a Dynal MPC-1 separator (Dynal, product #12001).
m 20 ml of pooled PCR reaction were mixed with 20 ml of Dynabeads M-280 in 20 mM tris-HCI (pH 7.4), 2 mM EDTA (pH 8.0), 2 M
NaCI. The tube was incubated at room temperature for 1 hour with mixing on a Denley Spiramix 5.
The Dynabeads were then washed four times with 20 ml of 10 ~s mM tris-HCI (pH 7.4), 1 mM EDTA (pH 8.0), 1 M NaCI. A fifth wash was performed in 20 ml of 1 x buffer M.
EcoRl-digestion was carried out for 1 hour at 37°C in 5 ml of 0.25 U/ul EcoRl in 1 x buffer M. The digest was then divided into ten 500 ul aliquot parts. Each aliquot was ethanol precipitated by the addition of 1 ~o ul of See DNA, 0.1 volume of 3 M sodium acetate (pH 5.2), and 2.5 volumes of ethanol. The precipitations were mixed and chilled to 0°C on ice for 30 minutes and then centrifugation at 20,000 maxRCF for 10 minutes. The pellets were rinsed with 70 % ethanol before dissolving in a total of 500 ~I of 1 x TE buffer.
~s A 1 ~~I aliquot of the EcoRl-released internal control DNA was electrophoresed on an 8 % polyacrylamide gel to confirm that the size of the purified DNA was in accordance with that expected.

High molecular weight genomic DNA digestion in the presence of the 4 by cutter internal control DNA with a dilution series of the enzymes Haelll, Msel, and Mbol High molecular weight human placental DNA was mixed with a dilution series of a mixture of the three set 1 4 by cutter restriction endonucleases - each at the same number of units. Aliquots were then removed and mixed with the EcoRl-released internal control DNA. 3.6 ~g of Human placental DNA (Sigma) was digested with 0.5 U/ul, 0.1 U/~I, 0.02 U/yl, 0.004 U/yl, and 0 U/ul each of Haelll, Mboi, and Msel in a total »~ volume of 36 ul. 0.4 yg of placental DNA and 1 ~I of EcoRl-released internal control DNA were digested with 0.5 U/~I, 0.1 U/~I, 0.02 U/~I, 0.004 U/~I, and 0 U/ul each of Haelll, Mbol, and Msel in a total volume of 4 pl.
Vistra Green (Amersham Pharmacia Biotech) staining was used to monitor the extent of internal control DNA digestion and high molecular weight DNA
~s digestion.
A premix of restriction enzymes (5 U/~I each enzyme), buffer, and BSA was prepared as described below:
Component ~I

50 U/~I Haelll i 1 20 U/~I Msel 2.5 25 U/~I Mbol 2 10x NEB buffer #2 1 10x BSA 1 Water 2.5 Total ~ 10 ?o Serial 5-fold dilutions of the premix were prepared in 1 x NEB
buffer #2 and 1 x BSA.
A 6x premix of human placental DNA, buffer, and BSA were also prepared as described below:

Component ~ 6x mix (~I) per reaction (~I) 1 mgiml human placental DNA i 24 , 4 10x NEB buffer #2 I 9.6 1.6 10x BSA i 9.6 1.6 ' Water ~ 52.8 , 8.8 This premix contained placental DNA at 0.25 mg/ml in 1x NEB buffer #2 and 1x BSA.
Human placental DNA and restriction mixes were then set up as follows:
', Component #1 ' Component #2 Tube 1 16 ul of placental DNA at 4 ul of each restriction 0.25 mg/ml in enzyme at 5 U/~I

1 x NEB buffer #2 and 1 in 1 x NEB buffer #2 and x BSA 1 x BSA

2 16 ul of placental DNA at 4 ul of each restriction 0.25 mg/ml ih enzyme at 1 U/~I

1 x NEB buffer #2 and 1 in 1 x NEB buffer #2 and x BSA 1 x BSA

3 j 16 yl of placental DNA at 4 ~I of each restriction 0.25 mg/ml in enzyme at 0.2 1 x NEB buffer #2 and 1 U/~I in 1 x NEB buffer #2 x BSA and 1 x BSA

4 16 ~I of placental DNA at 4 gl of each restriction 0.25 mg/ml in enzyme at 0.04 1 x NEB buffer #2 and 1 U/~I in 1 x NEB buffer #2 x BSA and 1 x BSA

j 5 16 ~I of placental DNA at 4 ul of 1 x NEB buffer #2 i 0.25 mg/ml in and 1 x BSA

1 x NEB buffer #2 and 1 i x BSA

2 ~I aliquots were then removed from tubes 1-5 and added to 1 pl of EcoRl-released internal control DNA and 1 ~I of 2x NEB buffer #2 to and 2x BSA to give samples 1-Sic. Samples 1-Sic were then overlaid with 50 ~I of mineral oil in order to prevent evaporation.
The remaining volume from tubes 1-5 was then added to 18 ~I of 1 x NEB buffer #2 and 1 x BSA.
All samples were finally incubated at 37°C overnight.

Restriction digests - final conditions Restriction digests 1-5 were therefore performed under the following conditions:
tube ~g humanU U U NEB ugiml placentalNaelll Msel Mbol buffer BSA
DNA #2 1 3.6 20 20 20 1x 100 2 3.6 4 4 4 1x 100 3 3.6 0.8 0.8 0.8 1x 100 4 3.6 0.16 0.16 0.16 1x 100 3.6 0 0 0 tx 100 Total volume = 36 ul Likewise, restriction digests 1-Sic were performed under the following conditions:
~o tube Ng humanul U U U NEB pg/ml placentalIC Haelll Msel Mbol buffer BSA
DNA DNA #2 1 0.4 1 2 2 2 1 x 100 2 0.4 1 0.4 0.4 0.4 1x 100 3 0.4 1 0.08 0.08 0.08 1 x 100 4 0.4 1 0.016 0.016 0.016 1x 100 5 0.4 1 0 0 0 1x 100 IC = EcoRl-released internal control DNA
Total volume = 4 pl is pl of digests 1-5 were each mixed with 3 ul of loading dye and 4 pl of digests 1-Sic were mixed with 1 ~I of loading dye. To sample number Sic, 180 ng of PCR molecular weight markers were added to serve as size standards. The band sizes for these markers are 50 bp, 100 bp, 200 bp, 300 bp, 400 bp, 500 bp, 525 bp, 700 bp, and 1000 bp. The digests were then electrophoresed on an 8 °o polyacrylamide gel in 1 x TBE as described below:
ml of 10x TBE
19.5 ml of 40 °~o acrylamide 19 ml of 2 °o methylene bisacrylamide 50.4 ml of water 1 ml of freshly prepared 10 % (w/v) ammonium persulphate 100 ul of TEMED
io A Cambridge Electrophoresis Ltd. vertical protein electrophoresis unit was used with 1 mm plate spacing. The samples were electrophoresed at 30 mA for 2 hours in 1x TBE. The gel was then stained for 30 minutes in 500 ml of 1 x TBE containing 50 ~I of Vistra Green. The is stained gel was imaged on a Fluorimager with the following settings: a 488 nm laser; a 570 DF 30 filter; a PMT setting of 700 V; 200 ~~m resolution;
and low sensitivity.
The results are shown in figure 3.
~o Example 2 TRSPA-2 analysis with the pNW33 BamHl, Hindlll, and Aflll matrix and the pNW33 Hindlll, Ncol, and Spel matrix probed with the 140 by BspEl to BamHl fragment from pNW33 ?s As an example of the case where the probe hybridizes to an arrayed PCR fragment with a different restriction site at each end, the pNW33 BamHl, Hindlll, and Aflll matrix (matrix #7) was probed with a PCR
product from the 140 by BspEl (466) to BamHl (605) restriction fragment within pNW33. In this example, the probe binds to a 204 by Hindlll to ~o BamHl PCR product derived from the 160 by Hindlll (445) to BamHl (605) restriction fragment within pNW33 (see below).

~~~rlll (4d5?
~-~~Bso~l (466) 140 bp~s, eamNl (605) 150 bp\ y BsrLl f755) 160 bp'y,h'indlll (975) 1'0 by ~ 7Ucol ('1085) 1 180 by pfJUw'33 (3554 bp) ~ -t- ~pel (7265) I
90 bp~
(not drawn to scale) .~p0 by ~ .1111 0455) 6sp ~l (9655) ,'.'~ h'indJll (1686!
~' h'indlll (7 77 5) As an example of the case where the probe hybridizes to an arrayed PCR fragment with the same restriction site at each end, the pNW33 Hindlll, Ncol, and Spel matrix (matrix #17) was probed with a PCR
product from the 140 by BspEl (466) to BamHl (605) restriction fragment within pNW33. In this example, the probe binds to a 514 by Hindlll to Hindlll PCR product derived from the 470 by Hindlll (445) to Hindlll (915) io restriction fragment within pNW33 (see above).
Oligonucleotides BamHl short PCR primer 5' TGTAACGACACATTGCTGGATACC 3' ~s Hindlll short PCR primer 5' ATATAACTCTCGCTCCTTGATAAC 3' Ncol short PCR primer 5' AGGCGTCTGAGGCTGCGGCTATGG 3' Spel short PCR primer 5' AACCCGTCGCGACGAGAGTCTAAG 3' Aflll short PCR primer 5' GATATACGTGATATATTTTGATTG 3' 2o BamHl adaptor 5' pGATCGGTATCCAGCAATGTGTCGTTACA 3' Hindlll adaptor 5' pAGCTGTTATCAAGGAGCGAGAGTTATAT 3' Ncol adaptor 5' pCATGCCATAGCCGCAGCCTCAGACGCCT 3' Spel adaptor 5' pCTAGCTTAGACTCTCGTCGCGACGGGTT 3' Aflll adaptor 5' pTTAACAATCAAAATATATCACGTATATC 3' BamHl long PCR primer 5' TGTAACGACACATTGCTGGATACCGATCC 3' Hindlll long PCR primer5' ATATAACTCTCGCTCCTTGATAACAGCTT 3' Ncol long PCR primer 5' AGGCGTCTGAGGCTGCGGCTATGGCATGG 3' Spel long PCR primer 5' AACCCGTCGCGACGAGAGTCTAAGCTAGT 3' Aflll long PCR primer 5' GATATACGTGATATATTTTGATTGTTAAG 3' Luc140down primer ~c~ 5' GCGCTAGGGATCCTTACTGGGACGAAGACGAA 3' Luc140up-bio primer 5' biotin-CGCAGCTGGTAATCCGGACGCCCGCGTCGAAGATGTT3' cs Restriction digestion of pNW33 Restriction digests of pNW33 were set up by combining the following:
~o Matrix #7 450 22 yg/ml pNW33 in 1 x NEB buffer #2 + 100 ug/ml ~I BSA

50 0.6 U/pl BamHl in 1 x NEB buffer #2 + 100 ~g/ml ~I BSA

50 0.6 U/~I Hindlll in 1x NEB buffer #2 + 100 ~I ug/ml BSA

~I 0.6 U/~I Aflll in 1x NEB buffer #2 + 100 ~g/ml ~I

Matrix #17 450 22 ~~g/ml pNW33 in 1 x NEB buffer #2 + 100 ~I ug/ml BSA

50 ~I 0.6 U/pl Hindlll in 1 x NEB buffer #2 + 100 ~g/ml BSA

50 ~I 0.6 U/~I Ncol in 1 x NEB buffer #2 + 100 pg/ml BSA

50 pl 0.6 U/~I Spel in 1 x NEB buffer #2 + 100 ~g/ml BSA

The samples were then vortex mixed, and incubated at 37°C
overnight.
Calf intestinal alkaline phosphatase (CIAP) digestion of pNW33 restriction digests 400 ul fractions of the 20 restriction digests (each containing 6.6 ~g of digested pNW33) were ethanol precipitated by the addition of 1 ~I
of See DNA, 0.1 volume of 3 M sodium acetate (pH 5.2), and 2.5 volumes of ethanol, chilling to -20°C, and then centrifugation. Pellets were rinsed m with 70 °o ethanol prior to re-dissolution in 20 ~I of 1 x CIAP
buffer containing 40 U of CIAP. The CIAP digests were carried out for 5 hours at 37°C and were then made up to 400 ~I with 1 x TE buffer.
Phenol extraction of pNW33 CIAP digests is The diluted digests were extracted with 400 NI of phenol and then ethanol precipitated as described above but with a 100 % ethanol wash after the 70 % ethanol wash. Samples were finally re-dissolved in 20 pl of TE buffer.
~o Annealing of short PCR primers to cognate adaptors Short PCR primers and their cognate adaptors were annealed by adding 1 pl of 200 pM short PCR primer to 1 NI of 200 NM cognate adaptor in 20 pl of 50 mM NaCI, 1x TE buffer. The mixed oligonucleotides were overlaid with 30 pl of light mineral oil and were then heated to 90°C
~s for 5 minutes followed by slow cooling to room temperature. The annealed short PCR primer / cognate adaptor complexes were then diluted with 1 ml of 1 x TE buffer and stored frozen at -20°C.
Ligation to annealed short PCR primers and cognate adaptors 0 1 pl of phenol extracted pNW33 digest was used per ligation.
Ligations were performed in 100 NI of 1 x ligase buffer containing 1 mM ATP

and 10 ul (100 U) of T4 DNA ligase. 20 NI aliquots from the 1 ml of annealed short PCR primer / cognate adaptor complexes were added according to the following table.
#7 BamHl l Hindlll l Aflll #17 Hindllll Ncoll Spel Ligation reactions were carried out for 24 hours at 16°C.
Samples were then diluted to 1 ml with TE buffer and stored at -20°C.
The diluted ligation reactions were then further diluted 1 in 10 and 10 NI was used as PCR template per 100 NI reaction.
io Restriction digestion of human placental DNA
1 U (~50 pg) of human placental DNA was digested overnight at 37°C with 100 U each of BamHl, BsrGl, Hindlll, Ncol, Spel, and Aflll in 400 pl of 1x NEB buffer #2 containing 100 pg/ml BSA.
IJ
CIAP digestion of human placental DNA restriction digest The digest was ethanol precipitated by the addition of 1 ~I of See DNA, 0.1 volume of 3 M sodium acetate (pH 5.2), and 2.5 volumes of ethanol, chilling to -20°C, and then centrifugation. The pellet was rinsed ~o with 70 % ethanol prior to re-dissolution in 50 pl of 1x CIAP buffer containing 100 U of CIAP. The CIAP digest was carried out for 5 hours at 37°C and was then made up to 400 NI with 1x TE buffer.
Phenol extraction of human placental DNA CIAP digest 2s The diluted CIAP digest was extracted with 400 pl of phenol and then ethanol precipitated as described above - again with a 100 ethanol wash after the 70 % ethanol wash. The sample was finally re-dissolved in 10 pl of TE buffer.

Ligation to annealed short PCR primers and cognate adaptors Short PCR primers were annealed to their cognate adaptors as described above.
The ligation to annealed short PCR primers and cognate adaptors was carried out in 100 pl of 1 x ligase buffer with 1 mM ATP. 100 U of T4 DNA ligase and 10 pl of each of the six short PCR primer / cognate adaptor complexes as above.
The ligation reaction was carried out for 24 hours at 16°C.
The sample was then diluted to 1 ml with 1 x TE buffer and stored at -20°C.
io 0.2 NI was used as PCR template per 100 pl reaction.
PCR amplification conditions An initial touch-down reaction was carried out in 50 pl of 1 x PCR buffer with all four dNTPs at 200 NM and Taq DNA polymerase at ~s 0.05 U/NI. Long PCR primers were used at 400 nM. 10 pl of pNW33 PCR
template was used per reaction and 0.2 pl of human placental DNA PCR
template was used per reaction. The samples were overlaid with 40 pl of light mineral oil and were touch-down thermocycled as described below:
~0 98°C for 1 min, 72°C for 2 min, 72°C for 5 min 98C for min, 69C for 2 min, 72C
1 for 5 min 98C for min, 66C for 2 min, 72C
1 for 5 min 98C for min, 63C for 2 min, 72C
1 for 5 min 98C for min, 60C for 2 min, 72C
1 for 5 min ?5 The main thermal cycling was then carried out in 100 ~I of 1x PCR buffer with all four dNTPs at 200 NM and Taq DNA polymerase at 0.05 U/NI. Long PCR primers were used at 400 nM. The samples were subjected to 20 cycles of 98°C for 1 min, 60°C for 2 min, 72°C for 5 min ~o and then 72°C for 10 min followed by chilling at 10°C and recovery from under oil.

WO 00/55364 PCT/GB00/00916 =

Arraying onto nylon membranes 1 ul aliquots from the PCR amplified samples were spotted onto Hybond N+ nylon membranes. The membranes were then transferred to a stack of three sheets of 3MM paper, saturated with 0.4 M NaOH, and incubated for 10 minutes. The NaOH was rinsed away in 2x SSC and the membranes were used directly for the pre-hybridization.
Probe synthesis ~c~ A PCR master mix was prepared as follows (all volumes are in ~I):
Master Mix Preparation 140 ~

W ater ~ 183.6 10x PCR buffer 30 2 mM dATP 15 2 mM dGTP

2 mM dTTP i 15 2 mM dCTP I 3 140 probe template I 1 Luc140up-bio 2.5 '- Luc140down 2.5 Taq DNA polymerase (5 2.4 U/pl) Total 270 Five PCR reactions were performed. For each reaction, 5 ~,I
of 33P-labelled dCTP was added to the PCR tube. The reactions were then made up to 50 ~I by the addition of 45 ~I of the master mix to each tube. Fach reaction was gently mixed.

PCR cycling parameters 94°C 2 min 50 °C 2 min 72°C 2 min 94°C 45 sec 25 cycles 50 °C 1 min 72°C 8 min 4°C hold ~o After thermal cycling, the five PCR amplifications were pooled. The 250 lul of labelled DNA was then used for the following magnetic bead purification procedure.
Capture of biotinylated PCR product and release of non-biotinylated > > strands All separations were carried out using a Dynal MPC-4 separator (Dynal, product #12004).
250 ~I of the pooled PCR was mixed with an equal volume of mg/ml streptavidin-coated colloidal Fe3O4 particles in 20 mM tris-HCI
(pH 7.4), 2 mM EDTA (pH 8.0), 2 M NaCI. The tube was incubated at room temperature for 1 hour with mixing on a Denley Orbital Mixer.
The streptavidin-coated colloidal Fe3O4 particles were then washed with 1 ml of 10 mM tris-HCI (pH 7.4), 1 mM EDTA (pH 8.0), 1 M
NaCI. Three more identical washes were performed.
2s The washed streptavidin-coated colloidal Fe3O4 particles were incubated in 500 ~I of 0.1 M NaOH for 10 minutes at room temperature. The supernatant was removed and added to 500 pl of 0.5 M

HEPES. The samples were then ethanol precipitated by the addition of 0.1 volume of 3 M sodium acetate (pH 5.2) and 2.5 volumes of ethanol. chilling to 0°C (on ice for 30 minutes), and then centrifugation at 20,000 maxRCF
for 10 minutes. The pellets were rinsed with 70 °i° ethanol before dissolving in 100 ~~I of TE buffer.
Hybridization Each membrane was placed in a 55 mm x 35 mm x 21 mm plastic box and 1.25 ml of pre-hybridization solution (5x SSC; Denhardt's ~o solution; 1 % SDS; 10 °,o dextran sulphate [Mw 500.000]; 0.3 tetrasodium pyrophosphate; 100 ug/ml denatured, sonicated DNA - pre-warmed to 65°C) was added. Each box was closed and incubated at 65°C
for 50 minutes on a rocking platform. The pre-hybridization solution was removed and replaced with hybridization solution (5x SSC; Denhardt's ~s solution; 1 % SDS; 10 % dextran sulphate [Mw 500,000); 0.3 tetrasodium pyrophosphate; 100 ~g/ml denatured, sonicated DNA -containing 5 yl of the appropriate 33P-labelled probe) and the box was incubated at 65°C for 3 hours on a rocking platform.
~o Washing The membranes were drained and transferred to 200 ml of 2x SSC, 0.1 % SDS at 68°C for 30 minutes. A further wash was carried out in 0.2x SSC, 0.1 % SDS at 71 °C for 30 minutes. The membranes were rinsed in 2x SSC at room temperature and laid out on blotting paper to ~s remove excess liquid. Once dry, the membrane was covered in Saran Wrap and exposed to a Kodak Phosphor Screen for 1 hour. The phosphor screen was subsequently imaged using a Molecular Dynamics Storm 860 Phosphorimager.
Results are shown in figures 4 and 5.
~o Example #3a A single cycle of inter-population perfectly matched duplex depletion wherein E.coli MutS protein is used to capture an A/A mismatch-containing duplex 'Affected' versus 'unaffected' (i.e. inter-population) mismatch-containing duplex selection can be achieved by: attaching a mismatch-binding protein to a solid support (or using the mismatch-binding protein in solution followed by subsequent solid-phase capture); taking denatured o> 'affected' DNA fragments and hybridizing these to denatured and biotinylated 'unaffected' DNA fragments; and capture of mismatch-containing duplex molecules with the mismatch-binding protein. Releasing the mismatch-containing duplex molecules (without strand denaturation), streptavidin capture and then release of the non-biotinylated strands will Is give only the desired species.
In this example, PCR fragments are prepared and used to demonstrate each of the individual steps for a single cycle of inter-population perfectly matched duplex depletion using E.coii MutS protein.
~o Clone insert design The clone inserts were constructed using standard cloning methodology well known to those skilled in the art and were inserted between the Aval site and EcoRl of pMOS8lue (Amersham Pharmacia Biotech).
~s The clone inserts contain a common 9 base pair internal core sequence in which a single nucleotide change or an insertion can be located. The internal core sequence is derived from codons 272-274 of human p53. These codons (GTG CGT GGT) correspond to a mutational hotspot found in lung and other types of cancer (R273L). For the design of ~o the clone inserts containing a mismatch (only one from a complete series of which is shown below), the nucleotide in position 5 of this core sequence is modified (GTGCXTGGT).
The internal core sequence is flanked by a random sequence - allowing the independent detection of the clone #1 and the clone #7 insert sequence in a mixed population of clone inserts.
s Clone #1 Mutant sequence (#1 M) ~'CCCGGGGGATCCTCGT""T""FmTGGGCCGAGTTTTGGTCCGTAGTGCTTGGTTAGATATGCTTAT
.. ' GGG'~CCCCm.~~GG.'~~1~JCTt-ll~ClA i.'T'.I:CC~GGCTCAA~CCAGGCtITCACGAAI..v, AAT~GTL-ITCjCGAATA
GTTCACAAAATCATCCTTGTACAGAATTC3' CAAGTGTTTTAGTAGGAACATGTCTTAAGS' Control sequence (#1 C) 5'CCCGGGGGATCCTCGTTTTATTGGGCCGAGTTTTGGTCCGTAGTGCATGGTTAGATATGCTTAT
.,'GGGCCCCCTAGGAGCAAAATAACCCGGCTCAAAACCAGGCATCACGTACCAATCTATACGAATA
?0 GTTCACAAAATCATCCTTGTACAGAATTC~' CAAGTGTTTTAGT~-.GGAACATGTCT'"AAGS ' Clone #7 ?_5 Control sequence (#7C) 5'CCCGGGTGTACACAAAAGTTTACCTGAAGAACGTGGGGGGTCGTGCCTGGTCTTGCGTCACCTG
3'GGGCCCACATGTGTTTTCAAATGGACTTCTTGCACCCCCCAGCACGGACCAGAACGCAGTGGAC
GTCTCAGGAGAGGGTCCCCATGGGAATTC3' CAGAGTCCTCTCCCAGGGGTACCCTTAAGS' Preparation of upper strand 5'-biotinylated and lower strand 5'-biotinylated #1 C double-stranded PCR product, and upper strand 5'-biotinylated and lower strand 5'-biotinylated #7C double-stranded PCR product Oligonucleotides BIOUPST2 5' bio-CTACTGATCGGATCCCCG 3' BIODOWN3 5' bio-AAACGACGGCCAGTGAAT 3' ~ o PCR reaction set-up (#1 C) #1 C in pMOSBIue at 2.50 ~g/ml 100 ~I

Water 8400 ~I

10x PCR buffer 1000 ~I

~s 50 ~M BIOUPST2 100 ~I

50 ~M BIODOWN3 100 pl mM dNTPs 2001 Taq DNA polymerase (5 U/~I) 100 ~I

~o PCR reaction set-up (#7C) #7C in pMOSBIue at 3.32 ~g/ml 100 ~I

Water 8400 ~I

10x PCR buffer 1000 ~I

~s 50 ~M BIOUPST2 100 ~I

50 ~M BIODOWN3 100 ~I

10 mM dNTPs 2001 Taq DNA polymerase (5 U/~I) 100 ~I

~o 96x 200 yl reactions were carried out for template #1 C and 96x 200 ul reactions were carried out for template #7C on a 96-well Perkin Elmer Cetus GeneAmp PCR System 9600 machine as described below:
95 °C, 5 minutes 1 cycle 95 °C. 1 minute 50 °C, 1 minute ' 30 cycles 72 °C. 1 minute i 72 °C, 5 minutes 1 cycle 4°C, hold The PCR products were pooled together and precipitated by adding 0.1 volumes of 3 M sodium acetate and 1 volume of isopropanol followed by centrifugation at 16.000 rpm for 30 minutes at 4°C (in a Centrikon T-2070 ultracentrifuge; swinging bucket Kontron rotor TST
~ c> 41.14). Pellets were washed with 14 ml of ethanol and centrifuged at 20,000 rpm for 30 minutes. Finally, the pellets were air-dried and resuspended in a total volume of 0.6 ml of TE buffer.
The 0.6 ml PCR sample was purified in twelve Microspin S-300 HR columns (50 ~I per column) following the manufacturer's protocol.
i ~ Briefly, the resin in the columns was resuspended by vortexing. Columns were centrifuged at 735 x g (3000 rpm in a microfuge) for 1 minute. The sample was then applied to the centre of the resin, being careful not to disturb the bed. The columns were centrifuged at 735 x g for 2 minutes and the flow-through containing the PCR product was collected. The ~o twelve eluted 50 ~I volumes were pooled together (pool 1 ). Columns were washed with 50 ~I of TE buffer and the eluted fractions were loaded onto a fresh S-300 HR column. Product yield and removal efficiency of the PCR
primers were analysed on a 1.5 % agarose gel.
?s Preparation of non-biotinylated #1 M single-stranded DNA and non-biotinylated #7C single-stranded DNA
Oligonucleotides BIOUPST2 5' bio-CTACTGATCGGATCCCCG 3' DOWN3 5' AAACGACGGCCAGTGAAT 3' PCR reaction set-up (#1 M) #1 M in pMOSBIue at 3.56 ug/ml 100 pl Water 8400 ~I

10x PCR buffer 1000 pl 50 ~M BIOUPST2 100 ~I

~s 50 ~M DOWN3 100 ~I

mM dNTPs 200p1 Taq DNA polymerase (5 U/~I) 100 ~I

PCR reaction set-up (#7C) #7C in pMOSBIue at 3.32 ~g/ml 100 pl Water 8400 ~I

10x PCR buffer 1000 ~I

50 pM BIOUPST2 100 pl ~s 50 ~M DOWN3 100 ~I

10 mM dNTPs 2001 Taq DNA polymerase (5 U/~I) 100 ~I

48x 200 pl reactions were carried out for template #1 M and :~0 48x 200 pl reactions were carried out for template #7C on a 96-well Perkin Elmer Cetus GeneAmp PCR System 9600 machine as described below:
95 °C, 5 minutes 1 cycle 95 °C. 1 minute ' 50 °C, 1 minute ' 30 cycles 72 °C, 1 minute 72 °C, 5 minutes 1 cycle 4°C, hold Capture of biotinylated PCR product strands and release of non-biotinylated strands for the preparation of non-biotinylated #1 M
single-stranded DNA and non-biotinylated #7C single-stranded DNA
ml of pooled #1 M and #7C PCR amplifications were each mixed with an equal volume of 4 mg/ml streptavidin-coated colloidal Fe304 particles taken up in 20 mM tris-HCI (pH 7.4), 2 mM EDTA (pH 8.0), 2 M
io NaCI. The tubes were incubated at room temperature for 60 minutes with mixing.
The streptavidin-coated colloidal Fe304 particles were then washed with 20 ml of 10 mM tris-HCI (pH 7.4), 1 mM EDTA (pH 8.0), 1 M
NaCI. Two more identical washes were performed.
~s The washed streptavidin-coated colloidal Fe304 particles were finally incubated in 800 ~I of 0.1 M NaOH for 10 minutes at room temperature. The supernatants were removed and added to 200 ~I of 2 M
HEPES (free acid). Samples were quantified by absorbance at 260 nm.
~o Denaturation and annealing Denaturation and annealing reactions were prepared by mixing:
''S

A = 400 ng of upper strand 5'-biotinylated and lower strand 5'-biotinylated #1 C double-stranded PCR product B = 400 ng of upper strand 5'-biotinylated and lower strand 5'-biotinylated #7C double-stranded PCR product C = 200 ng of non-biotinylated #1 M single-stranded DNA
(the lower of the two #1 M strands shown above) D = 200 ng of non-biotinylated #7C single-stranded DNA
(the lower of the two #7C strands shown above) Reannealing between the upper strand of A and the single-stranded C will therefore give rise to an A/A mismatch-containing duplex, 5'-biotinylated on the upper strand, for clone insert #1.
s Reannealing between the upper strand of B and the single-stranded D will therefore give rise to a perfectly matched duplex, 5'-biotinylated on the upper strand, for clone insert #7.
0.1 volumes of 1 M NaOH were then added, followed by incubation at room temperature for 10 minutes. 0.25 volumes of 2 M
~o HEPES (free acid) were finally added followed by incubation at 42°C
for 1 hour.
Samples were adjusted to 50 ~I and were made 1 x in PBS
and 1 mg/ml in BSA ready for reaction with MutS protein-coated magnetic beads.
is One 50 ~I sample (the pre-enrichment control, sample 6) was used directly for capture of biotinylated PCR product strands and release of non-biotinylated strands.
Mismatch capture 20 20 ~I of M2B2 MutS protein-coated magnetic particles (Genecheck, lot #20) were added to the annealed DNA above. Samples were incubated for 1 hour at room temperature with shaking.
Samples were then washed twice with 200 ~,I of ice-cold PBS.

Samples were finally eluted from the magnetic beads for 10 minutes at room temperature in 50 ul of the following:
Sample 1 M NaCI

Sample PBS

Sample 1 M urea Sample 1 % (w/v) SDS

Sample 10 mM NaOH

Capture of biotinylated PCR product strands and release of non-biotinylated strands All separations were carried out using an Amersham magnetic separator (RPN1682, batch #1).
50 pl of the eluates from the MutS protein-coated magnetic ~o beads and the pre-enrichment control (sample 6) were each mixed with an equal volume of 4 mg/ml streptavidin-coated colloidal Fe304 particles in 20 mM tris-HCI (pH 7.4), 2 mM EDTA (pH 8.0), 2 M NaCI. The tubes were incubated at room temperature for 30 minutes with regular mixing.
The streptavidin-coated colloidal Fe304 particles were then i s washed twice with 500 pl of 10 mM tris-HCI (pH 7.4), 1 mM EDTA (pH 8.0), 1 M NaCI at room temperature.
The washed streptavidin-coated colloidal Fe304 particles were incubated in 10 ~I of 0.1 M NaOH for 10 minutes at room temperature. The supernatant was removed and added to 2.5 ~I of 2 M
~o HEPES (free acid).
5 pl fractions were finally spotted onto Hybond N+ nylon membranes along with 200 ng, 20 ng, 2 ng, and 200 pg amounts of the following:
?5 C = non-biotinylated #1 M single-stranded DNA
(the lower of the two #1 M strands shown previously) D = non-biotinylated #7C single-stranded DNA
(the lower of the two #7C strands shown previously) Probe labelling Oligonucleotides #1 probe oligo 5' GGCCGAGTTTTGGTCCGTAG 3' #7 probe oligo 5' GTCTTGCGTCACCTGGTCTCAG 3' ~o Preparation of probes #1 probe oligo and #7 probe oligo were radioactively 5' end-labelled using T4 polynucleotide kinase as described below (all volumes are in pl):
#1 probe #7 probe #1 probe oligo (25 pmol) 1.25 -#7 probe oligo (25 pmol) - 1.25 10x PNK buffer 2.5 2.5 [y-33P~ ATP >92 TBq/mmol, >2500 Ci/mmol, 12.5 12.5 mBq/ml, 10 mCi/ml, (Amersham Pharmacia Biotech., AH9968, lot #B0006) T4 PNK (10 U/pl) 2.5 2.5 (Amersham International, E70031 Y, lot #201226) Water 6.25 6.25 The reactions were incubated at 37°C for 30 minutes and WO 00/55364 PCT/GB00/00916 =

then heated to 70°C for 5 minutes to denature the enzyme.
MicroSpin G-25 column purification Two G-25 columns (APB 27-5325-01, lot #9015325011 ) were resuspended by vortexing, and the bottom closures were snapped off as described in the manufacturer's instructions. A pre-spin of the columns was carried out for 1 minute at 730 maxRCF - 2670 rpm in a Hettich Zeutrifugen EBA 12 benchtop centrifuge - and the eluates were discarded.
The 25 ~I reactions were applied to each column, and the columns were m centrifuged for 2 minutes at 730 maxRCF. The eluates from the second spin were stored and used as probes.
Hybridization Each membrane was placed in a 55 mm x 35 mm x 21 mm > > plastic box and 2.5 ml of pre-hybridization solution (5x SSC; Denhardt's solution; 1 % SDS; 10 % dextran sulphate [MW 500,000]; 0.3 tetrasodium pyrophosphate; 100 ~g/ml denatured, sonicated DNA - pre-warmed to 42°C) was added. Each box was closed and incubated at 42°C
for 1 hour on a rocking platform. The pre-hybridization solution was =o removed and replaced with hybridization solution (5x SSC; Denhardt's solution; 1 % SDS; 10 % dextran sulphate [MW 500,000]; 0.3 tetrasodium pyrophosphate; 100 ~g/ml denatured, sonicated DNA -containing 2.5 ~I of the appropriate 33P-labelled probe) and the box was incubated at 42°C overnight on a rocking platform.
?5 Washing The membranes were drained and transferred to 200 ml of 2x SSC, 0.1 % SDS at 42°C for 10 minutes. A further wash was carried out in 0.2x SSC, 0.1 % SDS at 42°C for 10 minutes. The membranes were ~o rinsed in 2x SSC at room temperature and laid out on blotting paper to remove excess liquid. Once dry, the membrane was covered in Saran Wrap and exposed to a Kodak Phosphor Screen for 1 hour. The phosphor screen was subsequently imaged using a Molecular Dynamics Storm 860 Phosphorimager.
Dot blot layout Sample Sample 2 Sample Sample 4 Sample Sample pl fraction5 NI fraction5 gl fraction5 ul fraction5 pl fraction5 ul fraction 200 ng 20 ng 2 ng 200 pg - -#1 M ssDNA#1 M ssDNA #1 M ssDNA#1 M ssDNA

200 ng 20 ng 2 ng 200 pg ~ - -#7C ssDNA #7C ssDNA #7C ssDNA #7C ssDNA

ssDNA = single-stranded DNA
io Probe hybridisation results #1 probe ~ #7 probe I
i .: .
~ _~ _.:

i Signal intensities from the spots above were reported using ImageQuant 5.0 software (Molecular Dynamics), with the SumAboveBG
is figures being used after drawing a 6x3 grid over the array of spots.

#1 probe SumAboveBG signals Sample ~ Samp~e Sample 3 Sample Sample 5 Sample 200 ng 20 ng 2 ng 200 pg - -#1 M ssDNA #1 M ssDNA#1 M ssDNA #1 M ssDNA

I

200 ng 20 ng 2 ng 200 pg - -#7C ssDNA #7C ssDNA #7C ssDNA #7C ssDNA

#7 probe SumAboveBG signals Sample 1 Sample Sample 3 Sample Sample Sample 200 ng 20 ng 2 ng 200 pg - -#1 M ssDNA #1 M ssDNA#1 M ssDNA #1 M ssDNA

200 ng 20 ng 2 ng 200 pg - -#7C ssDNA #7C ssDNA #7C ssDNA #7C ssDNA

Recovery Elution conditions#1 signal #7 signal #1/#7 signal ratio (~o input ssDNA)(% input ssDNA) 1 M NaCI 21.6 % 10.0 % 2.2 PBS 3.7% 2.6% 1.4 1 M urea 26.4 % 17.7 % 1.5 1 % SDS i 59.9 % 23.0 % i 2.6 mM NaOH 70 .7 % 24.9 % I 2.8 to Recovery figures are plotted below for the various different elution conditions:

~~aG
,~O
s Example #3b A single cycle of inter-population perfectly matched duplex depletion wherein bacteriophage T4 endonuclease VIII protein containing a cleavage-inactivating N62D point mutation is used to capture an A/A
~o mismatch-containing duplex In this example, PCR fragments are again prepared and used to demonstrate each of the individual steps for a single cycle of inter-population perfectly matched duplex depletion using bacteriophage T4 endonuclease VIII protein containing a cleavage-inactivating N62D point i > mutation.
Clone design and DNA preparation Clone insert design was exactly as described in example #3a (see clone #1: mutant sequence (#1 M), control sequence (#1 C) and clone 20 #7: control sequence (#7C)).

Preparation of upper strand 5'-biotinylated and lower strand 5'-biotinylated #1 C double-stranded PCR product, and upper strand 5'-biotinylated and lower strand 5'-biotinylated #7C double-stranded PCR
product were as described in example #3a.
Preparation of non-biotinylated #1 M single-stranded DNA and non-biotinylated #7C single-stranded DNA were again as described in example #3a.
Denaturation and annealing ~c~ Denaturation and annealing reactions were prepared as described in example #3a.
Reannealing will therefore again give rise to an A/A
mismatch-containing duplex, 5'-biotinylated on the upper strand, for clone insert #1 and a perfectly matched duplex, 5'-biotinylated on the upper ~s strand, for clone insert #7.
One 50 pl sample (the pre-enrichment control) was adjusted to 100 pl with TE buffer and was used directly for capture of biotinylated PCR product strands and release of non-biotinylated strands.
~o Mismatch capture 50 pl samples of the annealing reaction were mixed with an equal volume of 200 mM sodium phosphate (pH 6.5), 100 mM KCI. 10 ~g of GST-tagged T4 endonuclease VII N62D mutant (obtained from Prof.
Borries Kemper, Univ. Cologne) were added to the annealed DNA and the 2s samples were incubated for 15 minutes at 16°C.
Samples were then mixed with 200 ~I of a 50 % slurry of Glutathione Sepharose 4B (Amersham Pharmacia Biotech, lot #279991 ) in 100 mM sodium phosphate (pH 6.5), 50 mM KCI. The tubes were incubated at 16°C for 30 minutes with regular mixing. The mixture was ~o then transferred to a spin column to separate solid from liquid phases.
Samples were finally eluted from the Glutathione Sepharose 4B matrix for 10 minutes at room temperature in 100 ~I of 10 mM reduced glutathione in 50 mM tris-HCI (pH 8.0).
Capture of biotinylated PCR product strands and release of non-biotinylated strands All separations were carried out using an Amersham magnetic separator (RPN1682, batch #1).
100 ~I of the eluate from the N62D T4 endonuclease VII
mismatch-capture reaction, and the pre-enrichment control were each ~o mixed with an equal volume of 4 mg/ml streptavidin-coated colloidal Fe304 particles in 20 mM tris-HCI (pH 7.4), 2 mM EDTA (pH 8.0), 2 M NaCI. The tubes were incubated at room temperature for 30 minutes with regular mixing.
The streptavidin-coated colloidal Fe304 particles were then ~s washed twice with 500 ~tl of 10 mM tris-HCI (pH 7.4), 1 mM EDTA (pH 8.0), 1 M NaCI at room temperature.
The washed streptavidin-coated colloidal Fe304 particles were incubated in 10 ~I of 0.1 M NaOH for 10 minutes at room temperature. The supernatant was removed and added to 2.5 ~I of 2 M
2o HEPES (free acid).
~I fractions were finally spotted onto Hybond N+ nylon membranes along with 200 ng, 20 ng, 2 ng, and 200 pg amounts of the following:
C = non-biotinylated #1 M single-stranded DNA
(the lower of the two #1 M strands shown previously) D = non-biotinylated #7C single-stranded DNA
(the lower of the two #7C strands shown previously) Probe labelling Probe labelling was exactly as described in example #3a.

Hybridization Each membrane was placed in a 55 mm x 35 mm x 21 mm plastic box and 2.5 ml of pre-hybridization solution (5x SSC; Denhardt's solution: 1 °,% SDS; 10 °o dextran sulphate [MW 500,000]; 0.3 tetrasodium pyrophosphate; 100 ~g/ml denatured. sonicated DNA - pre-warmed to 42°C) was added. Each box was closed and incubated at 42°C
for 10 minutes on a rocking platform. The pre-hybridization solution was removed and replaced with hybridization solution (5x SSC; Denhardt's m solution; 1 °,% SDS; 10 % dextran sulphate [MW 500.000]; 0.3 tetrasodium pyrophosphate: 100 pg/ml denatured, sonicated DNA -containing 2.5 pl of the appropriate 33P-labelled probe) and the box was incubated at 42°C overnight on a rocking platform.
is Washing The membranes were drained and transferred to 200 ml of 2x SSC, 0.1 % SDS at 42°C for 10 minutes. A further wash was carried out in 0.2x SSC, 0.1 % SDS at 42°C for 10 minutes. The membranes were rinsed in 2x SSC at room temperature and laid out on blotting paper to ~o remove excess liquid. Once dry, the membrane was covered in Saran Wrap and exposed to a Kodak Phosphor Screen for 1 hour. The phosphor screen was subsequently imaged using a Molecular Dynamics Storm 860 Phosphorimager.
Signal intensities from the spots above were again reported using ImageQuant 5.0 software (Molecular Dynamics), with the SumAboveBG figures being used after drawing a grid over the array of spots.

#1 probe SumAboveBG signals 200 ng 20 ng 2 ng 200 pg GSH eluted #1 M ssDNA #1 M ssDNA #1 M ssDNA#1 M ssDNAsample 200 ng 20 ng 2 ng 200 pg Pre-#7C ssDNA #7C ssDNA #7C ssDNA #7C ssDNA enrichment 229951 105343 16354 3632 control #7 probe SumAboveBG signals 200 ng 20 ng 2 ng 200 pg GSH eluted #1 M ssDNA #1 M ssDNA #1 M ssDNA#1 M ssDNAsample 200 ng 20 ng 2 ng 200 pg Pre-#7C ssDNA #7C ssDNA #7C ssDNA #7C ssDNA enrichment 3665967 22248096 203869 19406 control Recovery Elution conditions#1 signal #7 signal #1/#7 signal ratio (% input (% input ssDNA) ssDNA) mM reduced 14.5 % 10.0 % 1.45 glutathione in 50 mM

tris-HCI (pH
8.0) to The enrichment results are presented graphically below:

~ o Dedicated to the memory of Chris Griffin and Richard Beer #1 recovery (% input #7 recovery (°,o input ssDNA) ssDNA)

Claims (28)

-155-

1. A method of providing a mixture of DNA fragments enriched in fragments that are characteristic of a phenotype of interest, by providing affected DNA in fragmented form and providing unaffected DNA in fragmented form, which method comprises:
a) mixing the fragments of the affected DNA and the fragments of the unaffected DNA under hybridising conditions;
b) recovering a mixture of hybrids that contain mismatches;
c) recovering fragments of the affected DNA from the mixture of hybrids that contain mismatches;
and optionally repeating steps a), b) and c) one or more times.

2. The method of claim 1 wherein the affected DNA is pooled DNA of individuals who show the phenotype of interest, and the unaffected DNA is pooled DNA of individuals who do not show the phenotype of interest.

3. The method of claim 1, wherein the affected DNA is DNA of one individual who shows the phenotype of interest, and the unaffected DNA is pooled DNA of individuals who do not show the phenotype of interest.

4. The method of claim 1, wherein the affected DNA is DNA of one individual who shows the phenotype of interest, and the unaffected DNA is pooled DNA of a complete set of ancestors who do not show the phenotype of interest.

5. The method of claim 1, wherein the affected DNA is DNA
from cells of an individual that show the phenotype of interest, and the unaffected DNA is DNA from cells of the individual that do not show the phenotype of interest.

6. The method of any one of claims 1 to 5, wherein step b) is performed by use of a mismatch-binding protein.

7. The method of any one of claims 1 to 6, wherein either the fragments of the affected DNA or the fragments of the unaffected DNA are tagged by one member of a specific binding pair, and step c) is performed by using the other member of the specific binding pair.

8. The method of claim 7, wherein the fragments of the unaffected DNA are tagged with biotin, and step c) is performed by use of immobilised streptavidin.

9. The method of any one of claims 1 to 8, wherein the mixture of DNA fragments enriched in fragments that are characteristic of the phenotype of interest, is subjected to self-hybridisation followed by recovery of perfectly matched duplexes.

10. The method of any one of claims 1 to 9, wherein the mixture of DNA fragments enriched in fragments that are characteristic of the phenotype of interest, is mixed with an excess of the fragments of the affected DNA under hybridisation conditions, followed by recovery of perfectly matched duplexes.

11. The method of any one of claims 1 to 10, wherein each of the affected DNA and the unaffected DNA is provided in fragmented form by digestion with from 4 to 7 six-cutter restriction endonuclease enzymes together with from 0 to 5 four-cutter restriction endonuclease enzymes.

12. A mixture of DNA fragments enriched in fragments that are characteristic of the phenotype of interest, provided by the method of any one of claims 1 to 11.

13. A method of making a set of arrays of fragments of DNA of interest, which method comprises:
a) selecting, from a set of n restriction endonuclease enzymes, a subset of r restriction endonuclease enzymes;
b) digesting genomic DNA with the subset of r enzymes;
c) ligating to the resulting fragments restriction-enzyme-cutting-site-specific adapters with unique polymerase chain reaction amplifiable sequences;
d) splitting the resulting fragments into r2 aliquots;
e) amplifying each aliquot with two restriction-enzyme-specific primers of which one is tagged;
f) forming an array of the r2 aliquots of non-tagged amplimer strands; and g) repeating steps a) to f) using one or more different subsets of r restriction endonuclease enzymes.

14. A method of making a set of arrays of fragments of DNA of interest, which method comprises:
a) selecting, from a set of n restriction endonuclease enzymes, a subset of r restriction endonuclease enzymes;
b) digesting genomic DNA with the subset of r enzymes;
c) ligating to the resulting fragments restriction-enzyme-cutting-site-specific adapters with unique polymerase chain reaction amplifiable sequences;
d) splitting the resulting fragments into r2 aliquots;
e) amplifying each aliquot with two restriction-enzyme-specific primers;
f) forming an array of the r2 aliquots of the amplimer strands;
and g) repeating steps a) to f) using one or more different subsets of r restriction endonuclease enzymes.

15. The method of claim 13 or claim 14, wherein steps a to f) are repeated using each different subset of r restriction endonuclease enzymes to give (n!)/[(n-r)!r!] different arrays.

16. The method of any one of claims 13 to 15, wherein the n restriction endonuclease enzymes are selected from 4-cutters and 5-cutters and 6-cutters.

17. The method of any one of claims 13 to 16, wherein n is 3 to 10 and r is 2 to 4.

18. The method of claim 17, where n = 6 and r = 3.

19. A set of arrays of fragments of DNA of interest, which set results from performance of the method of any one of claims 13 to 18.

20. The set of arrays of claim 19, which set results from performance of the method of claim 13 and claim 14 and claim 15.

21. The set of arrays of claim 19 or claim 20, derived from a set of n = 6 six-cutter restriction endonuclease enzymes which are BamHI;
Bsr GI; Hind III; NcoI; SpeI; and AfIII.

22. The set of arrays of claim 19 or claim 20, derived from the set of n = 6 six-cutter restriction endonuclease enzymes which are EcoRI;
BspHI; BgIII; XbaI; Acc651; and ApaLI.

23. A nucleic acid characterisation method which comprises presenting to the set of arrays of any one of claims 19 to 22 a nucleic acid fragment of interest under hybridisation conditions, and observing a pattern of hybridisation.

24. The method of claim 23, wherein a plurality of nucleic acid fragments of interest are separately presented to the set of arrays, and the resulting patterns of hybridisation are compared.

25. The method of claim 24, wherein the plurality of nucleic acid fragments of interest are drawn from the mixture of DNA fragments, enriched in fragments that are characteristic of a phenotype of interest, of claim 13.

26. A method of identifying fragments of DNA that are characteristic of a phenotype of interest, which method comprises recovering, cloning and amplifying individual DNA fragments from the mixture of DNA fragments of claim 12, presenting the individual DNA
fragments to the set of arrays of any one of claims 19 to 22 under hybridisation conditions, observing a pattern of hybridisation generated by each individual DNA fragment, and subjecting to further investigation any two or more individual DNA fragments whose hybridisation patterns are similar or identical, or near to each other in a genome of interest.

27. A double-stranded DNA molecule having the sequence a-A-b-B...X-y-Y-z where A, B...X and Y are unique restriction sites for n different restriction endonuclease enzymes, and a, b...y, z denotes distances in base pairs, characterised in that each fragment, obtainable by cutting the DNA molecule by means of any one or more up to n of the restriction enzymes, has a different length from every other fragment.

28. The double-stranded DNA molecule of claim 27, wherein the following criteria are satisfied:
a) inter-fragment length differences are greater for larger fragments;
b) all possible fragments are unambiguousiy resolvable by electrophoresis from one another;
c) size gaps between bands comprising different numbers of inter-restriction-site units are larger than size gaps between bands comprising the same number of inter-restriction-site units;
d) the size gaps and size spread from the largest to the smallest fragment are electrophorectically compatible.