CA1099948A - Quantitative protein analysis by immunodiffusion - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE

Methods are described by which immunochemical procedures such as immunodiffusion and immunoelectro-phoresis can be used to provide quantitative measure-ment of the concentration of individual proteins in fluids such as serum, spinal fluid, tissue extracts and the like. That is accomplished typically by making quantitative optical measurements of the zones of pre-cipitation which are produced, in parallel with those of standard preparations containing known concentrations of the proteins to be determined; deriving from the measurements of the precipitation zones selected par-ameters; and comparing the selected values of the experimental and reference preparations. The zone measurements are preferably made by scanning the plates electronically, with suitable arithmetic manipulation of the resulting video values at a plurality of selected positions for each zone.

Description

QUArlTIT~'I`IVE PnO'r~IN Ail~LYS[S ~Y I~MUNODIY~USIOrl Thls invention relates generally to the quantita-tive measurement of protelns in mixtures, especially when only small amounts of materlal are availabie.
In view of the current rapid expansion of knowledge concerning the role of proteins in health and disease~
there is an increasing need for a general, rapid and rela-tively economical quantitative method for protein analysis of such fluids as serum, spinal fluid, tissue extracts and - /o the like.
The proteins occurring in such fluids are frequently identified by immunochemical procedures which depend upon precipitation of each protein by an antibody specific to the particular protein. The production of such specific antibodies is stimulated when foreign proteins (antigens) are introduced into a living body. Antisera can be pre-pared, containing known mixtures of such antibodies. By reacting a protein sample in vitro with such an antiserum and observing the resulting precipitation, or lack of precipitation, useful information may be obtained as to the types of proteins in the sample.
A primary ob~ect of the present invention is to obtain quantitative values for the concentration of one or more proteins in an antigen solution, utilizing immunodiffusion procedures which have previously been regarded as only qualitative techniques.

~6~99~

According to the invention a quantitative measure of the concentration oE a protein in an antigen sample is obtained after the sample has been sub]ected to immuno-diffusion with an antibody source containing an antibody specific to said protein, the protein and antibody diffus-ing into reactive contact generally parallel to an axis and through an area of a supporting medium initially free of both reactants to produce at least one elongated pre-cipitation zone of limited length, the method of obtaining such quantitative measure comprising scanning the zone by optical means to develop electrical signals responsive to precipitin concentration at a two-dimensional array of positions distributed partly within the zone and partly outside the zone, deriving electronically from said electrical signals the corresponding value of a selected parameter of the zone which varies in characteristic manner with said protein concentration, and comparing the resulting parameter value with a set of reference para-meter values derived correspondingly from reference zones produced by equivalent immunodiffusion of a plurality of reference antigen solutions containing respective known concentrations of said protein to provide a quantitative measure of the concentration of said protein in said sample.

.
The described collection of experimental data and the computations made with them can be carried out manually, if desired. They are also well adapted for semi-automatic data collection by optical and electronic scanning devices.
Also, the necessary data processing can be made fully automatic with the aid of a general purpose computer of moderate capacity.

~ 3999~3 In preferred form Or the invention, the proteins in the initial antigen mixture are first partlally fraction ated by causing them to migrate in one dimension at rates that differ characteristically among the various proteins.
Such selective migration may, for example, utilize slmple diffusion, electrophoresis, or more complex techniques such as chromatography without fundamentally altering the nature of the precipitin zones produced by the subsequent step of immllnodiffusion. In the lllustrative case of /0 electr~phoresis, differerlces of electrophoretic mobility between different proteins cause the proteins to become distributed in the direction of the elec-trical field in accordance with their mobility. Following such initial fractionation, the resulting essentially linear dlstribu-tion of proteins is brought into contact with the antiserum by relative movement in another dimension, typically by mutual diffusion in a suitable agar or agarose support medium. The precipitation zones of the respective pro-teins are then typically entirely separate and can be clearly disting~ished.
Useful separation of the precipitation zones of a plurality of distinct proteins is also attainable without an initial step of fractionation, if the antigen and anti-body are allowed to diffuse toward each other in a manner to form elongated precipitation zones of limited length extending transversely of the primary direction of diffu-sion. The mobilities for diffusion of different proteins, and/or of their antibodies, are ordinarily sufficiently different that such precipitation zones are clearly dis-30 tinguishable. Any overlapping that may occur is usually g~

limlted to portlons of the zones, the zone end points be-ing usually clearly separated. The quantltatlve proced-ures of the present inventlon are usefully applicable to lmmunodiffusion of that type.
Similarly, if antigen and antibody are placed into wells for immunodlffusion, and an electric rield is applled to accelerate the movement of the antigen and antibody to-ward each other, as in ~he procedure known as electro-immunodiffusion, the resulting precipitin zones retaln the /0 same basic fcrms as in absence of an electric field. The same is true if the lmmunodiffusion step of immunoelectro-phoresis is aided by a suitably directed electric field.
Accordingly, the term "lmmunodiffusion" in the present speclfication and claims refers to dlffusion with or without an acceleratlng electric field.

In the followlng description of certain illustra-tlve manners of carrying the invention into practice, reference will be made to the accompanying drawings in which:
o Fig. 1 is a schematic plan o~ an immunoelectro-phoresis plate, Fig. lA illustrating a typical distribu-tion of proteins following electrophoresis, and Figs. lB, lC and lD illustrating successive stages of the subsequent diffusion and immunoprecipitation reaction;
Flg. 2 is a schematic axial section representlng typical apparatus for measuring a slide;
Fig. 3 is a schematic drawing illustrating proper-ties of a precipitation zone relating to the invention;

4Ei Flg. Il ls a graph representing typical dependence of zone end polnts upon tlme;
Fig. 5 ls a graph representlng typlcal dependence of zone length upon tlme;
Fig. 6 is a graph representing typical dependence of tlme of inltial zone appearance upon protein concen-tration;
Flgs. 7 and 8 are graphs representing typlcal dependence of zone length upon time and ùpon protein ~O concentration;
Fig. 9 is a graph showing actual intenslty scans across a precipltation zone at two incubation times;
Fig. 10 is a graph representing intensity scans acro.ss precipitation zones formed by respectlve protein concentratlons;
Flg. 11 is a graph representing illustrative dependence of the parameter Iz upon time;
Fig. 12 ls a graph representlng illustrative dependence o~ the parameter Iz and of the parameter dIz/dt upon proteln concentration;
Fig. 13 is a schematic graph illustratlng deriva-tion of protein concentration values by known additions of the protein to the unknown sample; and Fig. 1~ is a schematic block diagram representing electronlc scanning apparatus for carrying out the in-vention.

9~8 Many aspects cr the invention are well illustrated by lts embodiment ln the preferred process Or lmmuno-electrophoresis, and the following description will empha-size that form of the invention, but wlthout implying any limitation of scope.
Immunoelectrophoresis 'ls well known as a quallta-tlve procedure, and many forms o~ apparatus ha~e been de-scribed for carrying it out, cliffering in detail rather than in principle. The electrophoresis and subsequent o diffusion and immunoreaction are typically performed in a single layer of gel from a fraction of a millimeter to several millimeters thick carried on an optically trans-parent sheet. A currently preferred supporting medium is agarose saturated with barbital buffer of approximately pH 8.6 and ionic strength 0.1. The active materials are typically inserted into wells, which may be cut from the gel coating or forned when the gel is molded on the carrier.
Fig. l shows schematically a plate 20 with a typi-cal arrangement of wells, comprising the two circularantigen wells 21 and 22, spaced equally on opposite sides of the elongated antibody well or trough 24 with axis 25, which extends parallel to the direction of electrophoresis.
With that geometric arrangement of wells two distinct antigen solutions~ or two specimens of the same solution, can be run slmultaneously against the same antibody solu-tion on each plate. If preferred, the capac~ty of each plate can be ~ultiplied, as by providing additional anti-body wells outward of the two illustrated antigen wells, with additional antl ~ ~ wells outward of them. Similarly, additlonal antigen wells may be ad~ed, spaced far enough from wells 21 and 22 in the direction of electrophoresis to prevent overlapping of the patterns.
The shaded areas 2~ and 28 in Flg. lA represent typical approximate distributions of four varieties of protein a, b, c and d from duplicate samples in the re-spective wells 21 and 22 following a period of electro-phoresis in the direction of the arrow 23. Althou~h all proteins usually migrate in the same direction through the liquid medium, the solvent itself tends to carry a net charge and to have a resultant flow or electroosmosis relative to the gel. Hence the proteins may have a net movement in either direction relative to the well.
Following termination of electrophoresis and addi-tion of antibody solution to well 24, precipitation zones are produced by mutual diffuslon of the proteins a, b, c and d of Fig. lA and the corresponding antibodies. Figs.
lB, lC and lD show typical zones at respective stages of development.
The precipitin arcs of unrelated antigens, reacting with the antibodies against them, form independently, and may cross when they are sufficiently close together, as shown in Fig. lD; whereas the arcs of antigens that are immunochemically related typically ~oin in a continuous reaction. The zone of precipitation d' in Figs. lC and lD indicates that the area d of Fig. lA actually contained two distinct proteins, illustrating the fact that even antigens which have identical electrophoretic mobility may form distinct zones of precipiiation. The zones d and d' .may alternatively be v~ewed as representing two distinct ~ ^~ ~
4~3 protelns that were lnitially placed in well 21 and were subjected to immunodirfusion without :Lnltlal electro-phoresls. The clear separation Or the respective zone ends would then be due to different rates of dlffusion of the proteins or of their antlbodies. In elther case, each such zone can be analyzed independently, using any or all of the methods of analysls to be descrlbed.
In accordance with the present lnvention, the zones of precipitatlon resulting from immunoelectrophoresis are /~ subjected to direct quantitative measurements. Such meas-urements may determine only the physical position on the plate of certain selected features of each precipitation zone of interest, or may include quantitative photooptical measurements of the light intensity. For both types of measurement, the time is noted and may be used as an inte-gral part of the observational data. If the incubation process is allowed to reach equilibrium, producing a nearly static zone configuration, the exact time of measurements becomes immaterial.
~ Position measurements on slide 20 can be made, for example, with the aid of a low power microscope. As rep-resented schematically in Fig. 2, plate 20 is placed over an adjustable opening 32 in the top of a light-tight box 30 with the lamps 36 and the dark backing 34 Or light absorbing material such as black velvet. The microscope 40 has the objective lens 413 the eye piece 42 and a set of cross hairs or other reference reticle at 43 in the focal plane. The microscope is mounted above light box 30 on a douole slide mechanism 45 with screw drive 46, O and with accurate scales, not explicitly shown, for readlng the microscope position in two coordinates. F'or clarity, only one coordlnate of the movement ls dlrectly lndicated in the drawing. It ls usually convenlent to select coordlnates havlng the x-axls, say, parallel to the direction of electrophoresis and to the length of the antlbody well, and having the origin of coordinates at or near the well axis 25, as indicated in Fig. lA
For measurements of the light intensity the micro-scope typically includes an oblique beam-splitting mirror /O 48 which sends part of the light to the eye piece while another part forms a real image in the plane of a dia-phragm 52. Diaphragm 52 then transmits to the photo-sensitive transducer 50 only radiation from the elemental area of plate 20 that coincides optically with the cross halr image. Transducer 50 is electrically connected to the amplifying circuit 54 and the meter 56. The latter may be observed visually and the value recorded manually, or the meter may embody means such as analogue-to-digital circuitry and printout mechanism for automatically record-O ing the light intensity in response to a command signal.Many changes can be made in the illustrative apparatus of Fig. 23 including, for example, replacing the dark field illumination by direct lighting or by top lighting such that light reflected from the zone is observed or measured.
Illustrative apparatus for making completely automated measurements is described below.
Fig. 3 illustrates schematically certain preferred precipitation zone features which are selected by the present invention for position measurement as the zone O develops. The horizontal line 61 at y = YAb represents the ad~accnt ~dge of the antibody well in the lmmuno-electrophoresis slide. The polnts E and F at the coordi-nates (xe, Ye) and (Xf,yf) represent the left and rlght end polnts of the elongated zone 60.
In additlon to the zone end points ~ an~ F, we have found it useful to measure coordinates of a number of intermediate points of each zone. The single point G at the coordinates (xg,yg) in Fig. 3 is illustrative of such points.
Since the zone has some width in the y direction, ~O the y coordinate of each intermediate point such as G
can be conveniently placed either at 63 at the leading edge of the zone closest to the antibody well, at 65 at the trailing edge of the zone, or at one or more points such as 64 within the zone, typically including the point of maximum intensity.
As the zone develops with increasing time of incubation, the points E, F and G progressively change both their relative and their absolute positions.
Illustrative values of xe and Xf as functions of time O are plotted in Fig. 4 for a typical protein concentra-tion. Such position values may be used directly for determining protein concentration, as by comparing the values of xe and Xf obtained with unknown solutions to corresponding values obtained with known concentrations of proteins at measured times. However, more reliable and accurate results are ordinarily obtainable by deriv-ing from such initial measurements one or more functlons, which will be referred to for convenience as parameters.

~ n important parameter of` the prec1p1tation zone employed by the present inventlon is the coordinate dif-ference xr - xe, which is a measure Or the length L or the precipitation zone at the particular time t of measurement.
The variation Or that parameter with time Or incubation is plotted in Fig. 5 for the typical data of Fig. 4.
Position parameters other than the zone length L
may be computed from measurements of the zone as it devel-ops. For example, the zone curvature and its variation ~O along the length of the zone are useful parameters, as well as providing information as to presence of protein abnorm-alities (see below). A rough measure of curvature for the zone arc as a whole, or for a selected zone segment, can be obtained by relatively simple comparisons of the x and y coordinates for three mutually spaced points on the zone axis. To obtain more accurate values of zone curvature, the zone axis is typically fitted approximately by a curve of the type y = f(x), where f(x) may represent any suitable function of x. The radius of curvature R is then given O by the general formula ~1 + (y') 337 (1) where y' and y" represent the first and second derivatives of y with respect to x.
A suitable illustrative function for curve fitting represents a parabola, typically with axis parallel to the y axis. Such a parabola may be expressed in either of the equivalent forms Y O + alx t a2x ~2a) and y = A(x _ B)2 t C (2b) 3 ~ where A = a2 al/2a2 C = a~ - al2/4a2 The values Or the constants in (2a) or (2b) can be found directly from the coordinates Or any three points Or the zone axis, or may be ritted by least squares or other known procedure to any desired nu~ber Or such points. The axls of symmetry of the parabola is at x = B, and the curve at that axis is a distance C from the x axis. Using the form-ula (1), the radius of curvature may be expressed as 1 ~[4A2(X _ B)2]3/2 (3) That radius has its maximum value Ro at the axis of sym-metry, where (3) reduces to ~O = 2A ~4) Any of the above quantities, which may be derived as indicated from as few as three points of the zone axis, may be employed as parameters in accordance with the invention.

Another parameter that is useful for determining protein concentra-tion in accordance with the invention is the time To of first appearance of the zone. That time is difficult to determine by direct observation. One aspect of the invention provides a practical way of ob-taining a reliable and reproducible value for the time of first appearance.
In Figs. 4 and 5 the solid lines represent typi-cal plots of direct experimental values. The figures also include extrapolations of the solid curves toward earlier times. The extrapolations are shown as dashed lines. The point 66 at which the extrapolated curves of Fig. 4 meet represents a time at which the zone must have had zero length. The extrapolated curve o~ Fig. 5 - llA -9~ .

intersects the time axls at the point 67, gl~ing an equlvalent ~rocedure ~or flnding To. Such extrapola-tion represents a reasonable and hlghly useful definition of the time of first appearance of the precipitation zone.
That method of determining To has the advantage that con-tinuous observation of the plate is not required.
Turning now to the quantitative determinations of the zone light lntensity, it has been discovered that a single light intensity reading does not usually provide 1~ a useful measure of protein concentration. That is pri-marily due to the variability of zone shape and speed of formation, and the tendency of the intensity to change as the zone expands.
On the other hand, we have found that the vari-ability of such factors can be largely compensated by tak-ing a series of intensity readings at suitably selected locations and treating them collectively to evaluate an intensity parameter. A preferred procedure is to take such readings at uniform intervals along a line extending O linearly across the zone of precipitation, typically in the y direction at a particular value of x. Such a series preferably includes several intensity values outside of the zone at each side. Those offset values are then aver-aged to provide a measure of the background intensity, which is subtracted from each of the intensity readings within the zone. The resulting adjusted intensity values are effectively summed, yielding essentially a linear integral of th~ intensity along a line crossing the zone at a selected value of x. ~e have found that such a linear lntensity sum Ix tends to increase with incubation 4~3 tlme ln a re~ular and reproduclble manncr~ the value at any given time increaslng with the concentration of the reacting protein over a wide range of experimental condi-tions.
Typical plots of the relative intensity observed during cross-zone scans are shown in Fig. 9. The two curves were plotted by semi-automatic apparatus of the type described below, scanning in the y direction at xg, the point of closest approach of the precipitation zones /0 to the ~ well. The peaks at 74 and 77 in ~ig. 9 are due to the zones formed on opposite sides of an anti-body well by identical protein samples in the two antigen wells of a plate slmilar to that of Fig. 1. The two small peaks at 75 and 76 are due to the respective edges of the antibody well, providing a convenient reference from which to measure distances such as D from selected zone points to that well. The parameter I defined above corresponds x essentially to the area under a peak such as those at 74 or 77 of Fig. 9.
Peaks 74 and 77 of Fig. 9 were made with identical samples of human serum albumin. They illustrate typical development of the precipitation zones between 2 hours of incubation (solid line curve 70) and 4 hours (dashed line curve 72). Although the zone position remains remark-ably stationary during the time between those two sets of measurements, the area of each peak ~rows appreciably.
The near identity of the peaks at 74 and 77 is noteworthy.
The two curves are offset vertically by an arbitrary dis-tance for clarity of illustration.

-13~

Flg. 10 ls a schematlc plot illustrating typical scans in the y dlrectlon on plates made with different concentrations of protein, all measured at substantlally the same tlme of lncubatlon. The graph brings out clearly the increaslng area of the indivldual peaks and the shlft of the entire zone toward the antibody ~iell wlth increas-ing protein concentration progresslng from peaks A to C.
Whereas the parameter Ix is highly useful for determining protein concentrations, still better results /0 are obtained from a multiple lntensity parameter, obtalned by summing or averaging such linear intensity sums at several different x values. A typical procedure is to compute linear sums at xg and at values spaced on each side of xg by a selected interval. Averaging or summing a uniform predetermined number of such linear sums reduces experimental error and improves the overall accuracy of the determination of protein concentration.
In accordance with another, generally preferred procedure, each time the plate is scanned the number of linear sums included in the computation is increased as the length of the precipitation zone increases. An illus-trative procedure of that type is to determine the linear sum of the intensity at xg and to continue to compute such sums on each side of xg until the value of the sum falls below a selected threshold. Addition of all the linear sums provides a parameter Iz which is essentially the inte~ral of the intensity of the zone of precipitation at the time of the scan. That approximation can be ob-tained as precisely as is des~red, within the limits of , resolution of the instrumentation, by reducing the x and y increments at which measurements are made. The value of Iz varies especially steeply as a function Or protein concentratlon over a wide range of experimental conditions.
That is because, as the concentration is lncreased, both the x and y dimensions of the zone increase, and the aver-age lntensity of the zone also tends to increase. That steeper dependence upon protein concentration makes the total intensity parameter especially effective as a cri-terion for determining the concentration.
After obtaining experimental values for one or more parameters for an antigen sample to be analyzed, those values are comparecl with suitable sets of standard values for the respective parameters, obtained under closely similar experimental conditions but with a series of pro-tein solutions containing respective known concentrations of the protein of interest. To obtain such an array of standard values, standard runs are carried out with such standard protein solutiGns, and measurements are made at corresponding points of the respective plates at succes-sive times as their incubation proceeds. The standard runs are preferably carried out with all conditions as nearly identi^al as possible to those of the experimental runs for which they are to provide reference values. In fact, a distinct set of reference values is preferably obtained for every group of experimental runs. However, for routine measurements for which the reference curve slopes are known from previcus experlence, even a single standard run may so~eti~es suffice.

Standard values of the selected parameter are de-rived from the results Or those standard measurements, typically for each concentration and at several times.
Each measured standard value Or the parameter is therefor consldered as a function of both concentration and time.
When lndlvidual linear intensity sums Ix are to be con-sidered separately, a full identification also requires specification of the value of x.
An advantage of using the time of first appearance /0 of the zone as a parameter is that standard values of To do not involve time as a variable. That is~ although the evaluation of To by the methods described requires measure-ments at a series of deflnite times, once To has been de-rived from those measurements the times of the respective measurements become immaterial. Thus, the values of To can be plotted as a function of protein concentration C, yielding a single standard curve. Such a curve is repre-sented schematically in Fig. 6, typically based on values obtained for respective concentrations by the extrapola-; ~ ~ tion method described in connection with Figs. 4 and 5.
Concentration values are directly readable from the curve of Fig. 6 for an unknown sample once its To has been meas-ured.
In the case of parameters such as L, Ix and Iz preparation of standard curves may be less direct. Since values depend upon the times at which the measurements are made, the series of reference standards must be pre-pared in such form tnat they cover a range of times.
It is not always feasible to measure data for all concen-3 ~ trations at the same moment. Each measured value is .

9~9~

therefore associated with lts tlme of' measurement, and the resulting standard values ror the respective protein concentrations are then plotted on separate curves as functions of time.
Fig. 7 shows a typical family of such curves in which standard values of the parameter for three concen-trations are plotted against kime. It is noted in pas-sing that the indicated extrapolation of those curves to L = O can provide standard va:Lues of To for plotting Fig.
f~ 6, or can provide experimental values of To for compari-son with Fig. 6. Vertical lines are drawn on Fig. 7 at a series of arbitrary times, shown as tl, t2 and t3.
- Their intersections with the curves then provide a set of L values for different concentrations, all corresponding to the same time. Each such set of L values is replotted as a separate curve as a function of concentration. The result is an array o~ curves, each showing L as a function of protein concentration for a particular time. Such an array is shown schematically in Fig. 8, and is found more convenient for comparison with an experimental parameter value than the plot against time of Fig. 7.
Standard values for comparison with experimental values of other parameters are typically obtained and treated in a manner analogous to that described for the parameter L.

We have discovered, however, that the total intensity parameter Iz, when measured for glven proteln concentratlon at successive times of zone development, lncreases linearly with the time. That linear relation is lllustrated in Fig.
119 which is a plot of Iz agalnst time for four solutions containing the indicated known concentrations Or the protein immunoglobulin. The indicated values of Iz were derived by a suitably programmed general purpose computer from intensity values measured automatically at a two-dimensional array of / G zone positions in the general manner to be described. The points shown in Fig. 11 were originally plotted automati-cally, and the straight-line curves were fitted to each set of points, by the computer. The figure has been redrawn manually and is reproduced at greatly reduced scale.
The llnear relatlon shown in Fig. 11 aids the pre-paration of standard or reference plots from which to read the protein concentration corresponding to an experimental value of the parameter Iz. One such curve, derived from the data of Fig. 11 and showing Iz as a function of concen-tration for the time 2.4 hours, is shown in Fig. 12 at 70.
The linear dependence of Iz upon time also means that the time derivative dIz/dt, or rate of change of Iz with time~ is constant. It thus provides a parameter having the practical advantage that it does not depend upon time.
Curve 72 in Fig. 12 illustrates typical behavior of that parameter as a function of protein concentration, each point representing the slope of one of the curves of Fig. 11 As already eY~plained in connection with the parameter To and Fig. 6, a single curve sucn as 72 of Fig. 12 can serve ~30 effectively as reference curve for the parameter dIz/dt.

A further aspect Or the lnvention prov1des improved accuracy, especially when the sample under study contains the protein or protelns of interest at 1 relatively low concentrations. The ~t~ ~ solutions are then preferably supplemented by adding known amounts of such proteins directly to aliquot portions of the sample itself, and running the resulting solutions in parallel with the original solution, and with standards containing definitely known concentrations of the protein.
~O Values are obtained for the desired parameter for the re-spective solutions all at a uniform time, uslng the above described technique for interpolating with respect to time if necessary. The parameter values P obtained for the regular standard solutions are plotted against protein concentration in the usual way, as indicated schematically by curve 80 in Fig. 13. Also plotted are the values of P
for the original antigen sample, and for the portions of that sample to which protein was added in known amount, indicated typically as the respective points h, i, j and k.
Those points are plotted relative to the horizontal concen-tration axis as if the solutions contained only the protein that was added. Thus the value h for the original sample is plotted at C = 0. Curve 81 is drawn through those points. With that illustrative arrangement of the data, the protein concentration in the original sample can be evaluated in several ways, which should gi~e essen~ially the same value. Thus, any desired number of those pro-cedures may be used, and the resulting values averaged.
First, horizontal pro~ection from point h to 3~ intersect curve 80 at h' yields the concentration value lndicated at Ch, ~ilich corresponds to the general compari-son procedure previously described. Eurther, slmllar pro~
~ection Or each value i, J and k to intersect curve 80 provides a measure of the concentration in terms of the lengths of the lines from i to i', j to J' and k to k', all of which lengths are theoretically equal. If the value of P at he is somewhat uncertain, for exa~ple be-cause the observed zone is faint, an average of all four intervals gives a more reliable value.
~ Another procedure has the advantage that it can not only improve the accuracy of an uncertain determina-tion of h, but can also provide an answer even if the original sample contains less than the threshold value of protein needed to produce a measurable precipitation zone.
In the latter case, the upper portion of curve 81 is drawn through the available points i, j and k, and is extrapolated to the P axisj thus providing a determination of P at h.
In thus pro~ecting curve 81, standard curve 80 is a helpful guide. For example, curve 80 is shifted bodily to the left in Fig. 13 till it represents as well as possible the points i, J and k. The intersection of the shifted curve 80 with the P axis then gives a good measure of point h, from which the concentration Ch is found. Also, further extrapolation of curve 81 to the negative C axis at -Ch gives a direct reading for the concentration, which should agree with the value Ch.
The described reliance upon the supplementary standard curve 81 has the great advantage of using the identical meclium for the experimental and reference solu-3 ~ tions. Especially when that medium is the human serum -of a partlcular patient, that advantage more than over-comes any possiblllty Or slight error in extrapolatlng the standard curve. By increaslng the number of dlfferent amounts of added proteln, the three~shown illustratlvely in Fig. 13 being merely illustrative, and by uslng several runs for each amount, the reliability of the extrapola-tlon can be increased almost without limit, and a good indication can be obtained of the experimental error that it may involve.
/0 The inventlon further provldes methods for the automated detection of abnormallties of` certaln proteins in an antigen sample. ~ormal and abnormal gamma globulin, for example, typlcally have sllghtly different ranges of electrophoretic mobility, but react with the same anti-body, leading to precipitation zones with irregular and typically unsymmetrical distribution of precipitin along the length of the zone. Such irregularity can be detected by comparin~ experimental values of the intensity param-eter Ix, for example, for different values of xO ~bser-vation of unsymmetrical or otherwise irregular varlation of such values indicates that an abnormality has been encountered; and that indica~ion is greatly strengthened if several independent determinations are carried out at each x value and if the computed experimental errors show the variation to be statistically significant.
That method may ~e considered as a special case of the general procedure of comparing the experimentally obtained values of different parameters ~Ihich tend to respond differently to the abnormality of interest. For 3 ~ some parameters ~lith such behavior it may be more effec-tive to compare the values Or proteln concentratlon thatcorrespond to the respective observed parameter values, rather than directly comparlng the parameter values them-selves. For example, the descrlbed lrregular lntensity distribution along the zone in presence of abnormal gamma globulin tends to cause the zone to appear sooner than normal~leading to hlgher values Or computed concentration;
whereas the total intensity parameter Iz tends to yield approximately equal concentration values for the normal J ~ and abnormal protelns. Thus significantly poor agreement between concentratlon values obtained with T and with I
o z indicates presence of abnormal protein.
A further parameter which responds sensitively to presence of abnarmal protein is the curvature of the longi-tudinal axis of the zone. In presence of abnormality the curvature tends to vary irregularly and unsymmetrically along the length of the zone, while the overall curvature tends to be less than normal.
Further advantages are nearly always obtainable by ~o employing multiple parameters which comprise specific func-tions of two OI' more parameters, such as the described position or intensity parameters. For example, the sum of the zone length L and an intensity parameter, each approp-riately weighted according to its experimental error of measurement, provides a new parameter which tends to give more reliable results than either one of its components when used alone.
Another advantageolis combination comprises a dif-ferentlal function of two parameters, one of which in-3 creases with lncreasing prcteln concentration, while the ~22-4~3 other decreases. 'rhus, a quotlent or a difference Or such parameters provides a multiple parameter wlth a steeper dependence upon concentratlon than elther Or lts compon-ents. The tlme To of first appearance of the zone is an example of a parameter having inverse relation to protein concentration, and is especial:Ly useful for constructing such quotients. The distance from the antlbody well to the zone at any deslred value of the x coordinate, for example the distance of closest approach at xg, also de-o pends lnversely upon the concentration and may be employedas a parameter in such quotients. It is generally pre-ferred to divide the parameter with direct dependence upon concentration by that with inverse dependence, so that the resulting multiple parameter has a direct rather than inverse net dependence.
Suitable reference values for comparison to experi-mentally determined multiple parameters of unknown samples are typically derived fro~ reference values obtained as already described for the respective component parameters.
Whereas the position and intensity measurements and the parameter derivations that have been described can be carrled out by direct visual and manual operations9 as has been indicated, an advantage of` the invention is that those operat~ons are particularly well adapted for partial or substantially complete automation.
An especially convenient and effective procedure for making the measurements required by the invention is to scan the slide by optical means, such as a television camera, a charge coupled scanning device or equivalent 3 0 means, for producing a video signal repre~enting the 4~

apparent brightness Or the scanned image at a two-dimensional array of elemental areas. That signal for each element is typically converted to digital form and electronically stored in association with signals repre-senting the x and y coordinates of the corresponding position on the plate and the time of observation. The entire array Or such position signals, or the slgnal for a designated position, can then be recovered from memory as needed for further processing. Also, any desired por-/ O tion of the slide image can be displayed either duringthe scanning operation or at a later time by means of a cathode ray tube or equivalent display device. Systems for scannin~ storing and reproducing an image, and for extracting a video signal in digital form for a selected image point are well known in the electronic art, and are available commercially in forms well adapted for the present purpose.
Fig. 14 represents schematically such apparatus in illustrative form, comprising the television camera 90 with O lens 91 for imaging the plate 20 on the photosensitive sur-face of the camera, the monitor cathode ray tube 92, the scan control apparatus 94 and the general purpose computer 100. Plate 20 can be imaged at any desired magniflcation by conventional ad~ustment of lens 91. A desired portion of the plate can be centered in the field by shifting its position on its illuminated support, such as light box 30 of Fig. 2, or by conventional electronic bias controls in the camera circuits.

More partlcularly, the scannlng movements in the scanning devlce and in the synchronized monitor catno~e ray tube, if used, are t~pically driven stepwise ln both coordinates, with steps so small that the movement aPpears vlrtually contlnuous on the screen. The image is thereby dlvided into elemental areas, which can be iaentified~
for example, by specifying thelr x and y coordinates.
For manual or semi-automatic operation, the monitor dis-play typically includes a cursor comprising an electroni-/o cally produced bright spot which designates on the screenthe elemental area from which the video sample is being extracted once each complete scan. T~e operator is pro-vided with manually controlled electronic switching means by which he can move the cursor about on the screen under visual control. Means may also be provided for directly placing the cursor at a desired point, typically in re-sponse to x and y signals in digital form constituting a command address. Such s1gnals may be derived under manual control, or supplied from a computer ln response to con-~2 ~ ventional program instruCtlOnS.
As illustratively shown in ~lg. 14, the counter 102 counts pulses received ~rom the ciock 103, producing on the multiple lines 104 digital signals which represent successive x coordinate values, say. Circuit 106 develops on the line 107 an analogue step voltage in response to those signals, wlth a step corresponding to each elemental x value. The dividing circuit 108 effectively counts the beam sweeps in Ihe x direction, producing on the multiple lines 109 digital sicnais representing the y coordinate ror each s~ ep. Clrcuit LlG develops on the llne 111 an analogue step volta~e ln response to that count, wlth a step corresponding to each elemental y value The step voltages on llnes 107 and 111 control the x and y sweep movements i~ camera 90 and tube 92, insuring thelr synchro-nizatlon. The video slgnal from camera 90 is supplied via the line 112 to monitor tube 92, reproducing on the tube screen the brightness variations of slide 20.
The selection circuit 118 typically comprises x / ~ and y counters which are selectively shiftable up or down either by a single count or in a continuing series of counts in response to manual movement of the ~oy stick indicated schematically at 119. The resulting di~ital signals, representing x and y coordinates of a point selected for exploration, are stored in register 117.
The comparison circuit 114 continuously compares the co-ordinates of that selected exploration address from reg-ister 117 with those of the scanning beams from lines 104 and 109. '~hen the scanning spot reaches the stored address, typically once during each complete scan, com-parison circuit 114 supplies an enabling pulse to the switching circuit 120. The video signal from camera 90 is thereby transmitted to analogue-to-digital circuit 122. The register 125 thus receives on the lines 123 a digital representation of the brightness of the scanned slide 20 at the selected exploration point. The enabling pulse from circuit 114 is supplied also to the cursor con-trol circuit 126, superimposing upon the video signal a beam intensi~`ying pulse ,lhich identifies that selected point on the monitor screen, as indicated schematically at 127.

- , ~ --r~egister 125 also receives the x an~ y coordinate signals from lines 115 and 116 and a digltal slgnal con~
tinuously representing the time, developed under clock control by the time circuit 128. All of those signals are typically stored in register 125 for delivery to com-puter 100 in response to unload signals which may be sup-plled either under manual control via the llne 129 from selector 118 or vla the llne 130 under control of the computer.
o Computer 100 is preferably provided with means for lndependently designating an exploration address and thus obtaining input informatlon for the spot of plate 20 that corresponds to speciflc requirements of any program under which it ls operating. Such exploration means may comprise circuitry basically similar to selector 118, register 117 and comparison circuit 1145 the selector operating in response to electronic signals which selectlvely indicate the required direction of movement of the exploration spot or the digital address of lts requlred position. Rather q~ ~ than duplicatlng such apparatus, Fig. 14 indicates computer control of selector 118 by electronic signals supplied via the multiple llnes 132, supplanting joy stick 119 when the switch 133 is shifted from manual position to automatic.
With such control available the computer is readily pro-grammed to carry out position and intensity measurements and parameter derivations of the general types that have been described.
For automatic scanning of zones, the plates of a set may be transferred in sequence from the incubation chamber to an accuratelv defined scanning position with sultable lllumlnation. Es~ecially when the optical scan-ning means is both light ln welght and compact, as ln the case of a charge coupled device, for example, it is gener-ally preferred to mount that scanning device on a platform that is movable ln two dimenslons over the statlonary array Or immunodiffusion or immunoelectrophoresls plates. That arrangemént facilitates use of the scannlng system as an aid to fill the antigen and antibody wells. For that pur-pose a digitally controlled dispenser is mounted in fixed /o but offset relation to the scanning devlce. Video signals from the scanner are supplied to the computer~ which is suitably programmed to recognize the shapes of the indi-vidual wells, or to identify machine readable symbols associated with them. As the platform carrying scanner and dispenser is moved under computer control over the plates~ it can then readily be stopped when the scanner axis is located accurately over a selected antigen well.
The platform may then be moved a fixed increment to center the dispensing tip over the well so that the correct charge is dispensed accurately into the well.
When all wells are filled in this manner with the appropriate antigen solutions~ with suitable washing of the dispenser between operations, eiectrophoresis is initlated. A similar loading operation is typically carried out to charge the antibody wells, either after completion of the electrophoresis, or immediately after filling the antigen wells if electrophoresis is omitted.
After the desired time for diffuslon the same scanning device is moved successively to all positions ~here zones O Or precipitat:Lon are to be scanr.ed throughout the array of plates.

The scanrling capabillty o~ the scannin~ ~evice ls preferably used also to scan the array Or plates berore the wells are loaded. The computer ls programmed to com-pare the observed positions Or the antigen and antibody wells, recording any departure from uniformity of dimen-sions or relative positions among the plates of the array.
Any plate or individual well lound to depart too far from standard can then be automatically omitted during the load-ing operation; and smaller deviations can be compensated / ~ by automatically applying suitable corrections to the parameter values that are ultimately derived by the com-puter.
Initial zone scanning procedure typically comprises request by the computer for video information at successive exploration addresses with a selected x value and with y values shifting progressively by a specified interval over the region in which zone formation is anticipated. The `video intensity values received from register 125 are stored with x, y and t data for each explored point. After ~? ~ each such y-scan, the x coordinate is shifted by a speci-fied interval and a similar y-scan is per~ormed and the re-sults recorded, until the entire area of interest has been covered. Each received intensity value is typically com-pared with the previously received and stored values. If the observed intensity variations exceed a set threshold characteristic of a background area, indicating presence of a precipitatlon zone, each intensity value within the zone is so designated in memory. Also, I~r each y-scan, the intensity ,mâximum as the zone is cros~ed is identified 3 ~ and recorded, establlshing the zone axis n terms OI' a 9~

series of y values at uniforrllly spacecl x values. 'l'he end points Or each zone are identified typically as the terml-nal x values in each direction at which an intensity maxi-mum was identifled. If more precise location of the end points is desired, the computer is instructed to perform further scans in a defined region about each end point with reduced x and y intervals.
With the existing precipitation zones so located and recorded, straightforward arithmetic operations are performed on the stored data, yielding the described in-tensity sum paranl.eter Ix for each cross-scan of a zone in the y direction; and addition of those values gives the total intensity parameter Iz for the zone. Subtrac-tion of the x value~ at the zone ends gives the parameter L. Additional parameters may be obtained by appropriate computation as desired.
The aescribed measurements are preferably carried out as a unitary operation on a complete set of plates that includes one or more unknown samples and also a set of related standard solutions of the type discussed above and sufficiently complete to permit evaluation of the unknown samples. After each scan of such a plate set, the computer is preferabiy directed to derive concentra-tlon values for each of the protelns for which precipita~
tion zones were found. If, as is preferred, multiple standard runs are included, the prcbable experimental error can be derlved for each calculated conc2ntration value. Computer routines for sucn calculatiGns are well known. Several independerlt determinations are preferably made of the concentration of each protein in the sample, typically by re~erence to d1frerent parameters; an average is then computed, with each value weighted in the usual way according to its probable error.
If the computed probable error for that average is within the specified range for the particular sample under study, no further measurements are needed. The computer then produces a conventional printout or other record of the final results, together with as much of the original data as may be requested by the program.
For example, the physician for whom the analysis is be-ing carried out may require a complete copy of all the data on magnetic tape or the like for possible future reference. Also, photographs of the immunodiffusion plate can be taken, either directly or using the monitor tube.
Ordinarily, one cycle of scanning does not permit an adequate quantitative determination Or a protein unless it is present in high concentration. Aft~r a suitable length of supplementary incubation, which may vary from a few minutes, say, to an hour or more, the above described scanning process is repeated, typically f~r the entire set of test plates and the corresponding stan~ards. The com-puter is typically instructed to explore plate areas where zones were expected but not found during ~he previous scan;
and also areas corresponding to the previcusly found and recorded precipitation zones, preferably ~llowing for specified expansion or movement of those ~ones ~hich are set into the program on the basis of prev-lous experience.
Thus, the computer operatior.s can maintain a continuity ~ ~\

Or treatment of the respective protein zones between one scan and the next.
Occasionally two precipltation zones due to dis-tinct proteins are so close together that their normal growth ultimately produces over-lap. For zones produced by immunoelectrophoresis, such overlap normally occurs at or near the zone ends, leading to crossing of the zones as shown typically for zones a and b of Fig. lD. When lmmunodiffusion is carried out without initial fraction-~O ation, the overlap tends to be confined to the zone centerportions, as when zones d and d' of that figure grow to-gether. The computer is typically programmed to anticipate such a zone overlap, to recognize it when it has occurred, and then to make suitable modifications in the procedure used for deriving the various parameters.
When the proteins under study are such that over-laps are expected, each plate is typically scanned at an early stage of zone development before zone overlaps have occurred (~ig. lB). The zone axis and end points can then be located without ambiguity. The computer is typically programmed as part of the regular processing of each scan to check for actual overlaps, for example by comparing the (x,y) coordina~es for each measured point of a zone axis with those for the adjacent zones, coincidence within a specified threshold indicating an overlap. All scans preferably also include a check for potential overlaps.
For example, the computer extrapolates each zone axis be-yond the observed end points ar.d compares the extrapolated axis points of adJacent zones. Such axis extrapolation ~3~ may comprise a simple linear extension ln khe direction of 4~

the axls slope near each en~ ltelnatively> the observcd zone axls may be fltted by a parabola or othcr curve, whlch can then be extrapolated accurately.
Zone overlap can also be detected by computlng the rate of change of slope of the zone axls, or the radius of curvature Or that axis, at a series Or points along the zone. Any departure from the normally smooth variation of those functions indicates that the zone may be made up of two or more overlapplng zones.
When an actual overlap is found on one half of a zone, sufficiently accurate compensation can often be made by slmply replacing the observed intensity values in the area of overlap with the values observed at the correspond-ing points of the other half of the zone. Such correspond-ence between two points for zones due to immunoelectro-phoresis is typically defined as equal x spacing on oppo-site sides of xg of the point of closest approach to the antibody well; and for zones due to direct immunodiffusion is defined as equal y spacing on opposite sides of the zone axis. The axis within the area of overlap can usually be located by extrapolation.
- More elaborate and accurate compensation procedures are also available, if desired. For example, the computer is instructed to adjust the measurement at the selected symmetrically placed point of the unaffected half of the zone to take account of any actual lack of symmetry of the zone. That unsymmetry can be determined, for example, by comparing the values that were obtained for the two points during a previous scan prior to the overlap, and 3 ~ applying the ratio of those values as a correction factor.

A fUl'thCr lllustrltlve colllperlsatlorl pr~occ~ure takes account also Or the lntenslty value actually meas-ured at the reglon of overlap, which ls due, of course~
partly to one zone and partly to the other. The computer is instructed to divide that observed value at each point of overlap ln an approprlate ratlo. A suitable approxi-mate ratio may be obtained by comparing measured values for the two zones at the respective symmetrically placed points already described, either with or without the ~U described adjustment of those values.
Whenever possible, it is preferred to ernploy two or more dlstinct computation procedures for compensating for areas of overlap, averaging the results and deter-mining the probable error. It is emphasized, however, that the region of overlap is ordinarily a small fraction of the entire area of a zone. Hence even quite appreciable errors in approximating the true value within the overlap still may not affect the final result significantly.
- Also, for each of the described types of overlap, it is usually possible to make the determination of pro-tein concentration in terms of parameters that are defined by portions of the zone not affected by the overlap. Thus, the measurements preferably employed are those near the zone center for processes similar to immunoelectrophoresis, and those near the zone ends for processes similar to direct immunodiffusion.
It will be recognized by ~hose skilled in the art that many particulars of the described procedures can be replaced by their obvio!ls equivalents Wit}lCUt departing 3 0 fro~ the proper scope of the present invention. For B

example, values ror those parameters that do no~ require a series Or rneasurements at successlve lncubation times can be obtained from measurements made after the incuba-tion process has been allowed to reach substantial equili-brlum. Such measurements have the advantage that the zone configurations are virtually static, and the exact time of measurement is therefore not crucial. As a further example, at an desired stage of the incubation the zones of precipitation can be stained in known manner with an appropriate dye, and zone measurements can then be made using selectively transmitted light. That method of meas-urement is usually most useful after equilibrium has been reached, since it is then not necessary to assign a pre-cise time of measurement.
In summary, methods have been described by which immunodiffusion, immunoelectrophoresis and analagous processes can be used to provide truly quantitative measurement Or the concentrations of individual proteins in biological fluids. Actual use of` the described methods has demonstrated that protein concentrations can be deter-mined within five percent or better. The invention thus provides an effective and convenient method for making a wide variety of experimental and diagnostic determi-nations.

Claims (34)

1. The method of obtaining a quantitative measure of the concentration of a protein in an antigen sample which has been subjected to immunodiffusion with an antibody source containing an antibody specific to said protein, the protein and antibody diffusing into reactive contact generally parallel to an axis and through an area of a supporting medium initially free of both reactants to produce at least one elongated precipitation zone of limited length; said method comprising scanning the zone by optical means to develop electrical signals responsive to precipitin concentration at a two-dimensional array of positions distributed partly within the zone and partly outside the zone, deriving electronically from said electrical signals the corresponding value of a selected parameter of the zone which varies in charac-teristic manner with said protein concen-tration, and comparing the resulting parameter value with a set of reference parameter values derived correspondingly from reference zones produced by equivalent immunodiffusion of a plurality of reference antigen solutions containing respective known concentrations of said pro-tein to provide a quantitative measure of he concentration of said protein in said sample.

2. The method according to claim 1 wherein, prior to said immunodiffusion, said antigen sample has been subjected to selective protein migration in a direction generally per-pendicular to said axis.

3. Method according to claim 1 wherein said parameter comprises the distance between the ends of said pre-cipitation zone.

4. Method according to claim 2 wherein said parameter comprises the distance between the ends of said pre-cipitation zone.

5. Method according to claim 3 or claim 4 wherein said step of scanning the zone is carried out transversely of the zone along a plurality of mutually spaced parallel scan paths which include scan paths crossing the zone and scan paths beyond the respective ends of the zone, and said step of deriving a parameter value comprises comparing said electrical signals to select scan paths adjacent the respective zone ends and determining the mutual separation of the selected scan paths.

6. Method according to claim 1 or 2 wherein said para-meter comprises the integrated zone intensity, said electrical signals represent light intensity at the respective positions of said array, and the step of deriving a parameter value comprises summation of signal values at positions within the zone corrected for values at generally adjacent positions outside the zone.

7. The method according to claim 1 or 2 wherein said parameter comprises the time of initial appearance of the zone.

8. The method according to claim 1 wherein said parameter comprises the time of initial appearance of the zone, said step of scanning the zone is repeated at a plurality of selected times as the zone develops, and said step of deriving said parameter value comprises deriving from the electrical signals for each said scan a value of a second parameter which varies in characteristic manner as a function of time and has a known value at the time of initial appearance of the zone, and extrapolating the resulting values of said second parameter back in time to said known value.

9. The method according to claim 2 wherein said parameter comprises the time of initial appearance of the zone, said step of scanning the zone is repeated at a plurality of selected times as the zone develops, and said step of deriving said parameter value comprises deriving from the electrical signals for each said scan a value of a second parameter which varies in characteristic manner as a function of time and has a known value at the time of initial appearance of the zone, and extrapolating the resulting values of said second parameter back in time to said known value.

10. The method according to claim 8 or claim 9 wherein said second parameter comprises the distance between the ends of the precipitation zone.

11. The method according to claim 8 or claim 9 wherein said second parameter comprises the integrated zone intensity.

12. Method according to claim 1 or 2 wherein said step of scanning the zone to develop electrical signals comprises forming an optical image of the zone at the image surface of a video camera tube, representing the intensity of the image at each said position as a digital representation, and storing in a memory associated with a digital computer the digital representations and digital addresses iden-tifying the respective image positions.

13. The method according to claim 1 or 2 including producing on the screen of a cathode ray tube an image responsive to said electrical signals and representing said zone.

14. The method according to claim 1 or 2 wherein said step of scanning the zone is repeated at a plurality of times as the zone develops, and said electrical signals rep-resent light intensity measurements at a plurality of positions, x and y coordinate values at the respective positions and time values corresponding to the respective measurements.

15. The method according to claim 1 wherein said step of deriving a parameter signal includes deriving a first signal which increases with increasing protein concen-tration, deriving a second signal which decreases with increasing protein concentration, and deriving as a combined parameter signal a third signal which represents a differential combination of said first and second signals.

16. The method according to claim 2 wherein said step of deriving a parameter signal includes deriving a first signal which increases with increasing protein concen-tration, deriving a second signal which decreases with increasing protein concentration, and deriving as a combined parameter signal a third signal which represents a differential combination of said first and second signals.

17. The method according to claim 15 or claim 16 wherein said combined parameter signal represents the ratio of said first and second signals.

18. The method according to claim 1 or 2 wherein said array of positions include at least one position within said zone on one side thereof and in a region of overlap with another zone, and said step of scanning the zone to develop electrical signals includes developing an elec-trical signal at a second position which is correspondingly placed on the other side of the first said zone and in a region free of overlap, and employing said second position signal for deriving said electrical signal for said one position.

19. The method according to claim 1 or 2 wherein respect-ive portions of said antigen sample have been enriched by adding different known quantities of said protein prior to said immunodiffusion of each sample portion, said method including subjecting the resulting zone for each sample portion to said steps of scanning, deriving parameter values and comparing the resulting parameter values with said set of reference parameter values.

20. The method according to claim 1 or 2 wherein respect-ive portions of said antigen sample have been enriched by adding different known quantities of said protein prior to said immunodiffusion of each sample portion, said method including subjecting the resulting zone for each sample portion to said steps of scanning and deriving parameter values, and effectively plotting said resulting parameter values for the respective sample portions against the protein concentrations added and extrapolating the resulting curve to obtain a value for said protein concentration in the original sample.

21. The method according to claim 1 or 2 including the step of staining the zone of precipitation with a dye prior to said step of scanning the zone.

22. Method according to claim 1 including the steps of deriving electronically from said electrical signals the corresponding value of a further parameter of the zone which differs characteristically from the first said parameter in presence of certain abnormalities of said protein, and comparing the values of the first said parameter and of said further parameter to detect presence of such abnormality.

23. Method according to claim 2 including the steps of deriving electronically from said electrical signals the corresponding value of a further parameter of the zone which differs characateristically from the first said parameter in presence of certain abnormalities of said protein, and comparing the values of the first said parameter and of said further parameter to detect presence of such abnormality.

24. Method according to claim 22 or claim 23 wherein one of said parameters comprises essentially the integrated zone intensity and the other said parameter compises the time of initial appearance of the zone, and the parameter values are compared in terms of the respective protein concentrations derived from them.

25. Method according to claim 22 or claim 23 wherein said one parameter and said further parameter comprise intensity sums extending across the zone at respective spaced positions along the zone length.

26. Method according to claim 1 including the steps of deriving electronically from said electrical signals the corresponding values, at a plurality of positions along the zone, of a zone parameter which is responsive to irregular or unsymmetrical precipitate distribution along the zone, and comparing said parameter values as a function of the positions along the zone length to detect presence of protein abnormality.

27. Method according to claim 2 including the steps of deriving electronically from said electrical signals the corresponding values, at a plurality of positions along the zone, of a zone parameter which is responsive to irregular or unsymmetrical precipitate distribution along the zone, and comparing said parameter values as a function of the positions along the zone length to detect presence of protein abnormality.

28. Method according to claim 26 or claim 27 wherein said parameter values represent the zone curvatures at a plurality of zone segments.

29. Method according to claim 26 or claim 27 wherein each said parameter value represents a summation of signal values at a plurality of positions distributed trans-versely of the zone.

30. Apparatus for obtaining a quantitative measure of the concentration of an antigen in an antigen containing sample which has been subjected to immunodiffusion with an antibody specific to said antigen, the antigen and antibody diffusing through a supporting medium from respective sources mutually spaced along an axis and reacting to form at least one elongated precipitation zone of limited length; said apparatus comprising means for automatically scanning the supporting medium to develop electrical signals representing light intensity at respective positions of a two-dimensional array dis-tributed partly within the zone and partly outside the zone, each light representing signal being associated with electrical position representing signals identifying the corresponding position in said array, and means responsive jointly to position representing signals and the corresponding light representing signals for deriving the value of a selected parameter of the zone which varies in characteristic manner with said antigen concentration.

31. Apparatus according to claim 30 including means for causing said scanning means to repeat said scanning action intermittently, and means for associating electrical time signals with said light representing signals.

32. Apparatus for obtaining a quantitative measure of the concentration of an antigen in an antigen containing sample by subjecting the sample to immunodiffusion with an antibody source containing an antibody specific to said antigen, said reactants diffusing in a supporting medium from respective antigen and antibody wells to produce a precipitation zone; said apparatus comprising photoresponsive means for scanning the wells and the supporting medium to develop electrical signals repres-enting light intensity at respective positions, each light representing signal being associated with electrical posi-tion representing signals identifying the corresponding position, dispensing means for dispensing antigen and antibody containing solutions selectively to the respective wells in response to electrical command signals, and control means responsive jointly to the light representing sig-nals and the position representing signals and having one mode of operation for producing operative position of the dispensing means relative to each individual well and for supplying a command signal to the dispensing means to supply that well with a selected reagent;
said control means having another mode of operation for deriving electronically from the light representing signals and the position representing signals the corres-ponding value of a selected zone parameter which varies in characteristic manner with said antigen concentration.

33. Apparatus according to claim 32, said control means, when in said one mode, including means for deriving sig-nals representing the position of each well, and being responsive to such well position signals for producing operative position of the dispensing means.

34. Apparatus according to claim 32 including means associated with said control means, when in said one mode, for comparing the relative positions of said wells with predetermined standard relative well positions to detect deviations therefrom, and means associated with said control means, when in said other mode, for compensating for such deviations during said parameter derivation.