WO2011073402A1

WO2011073402A1 - Transcript variants of vnn1 and slc39a14

Info

Publication number: WO2011073402A1
Application number: PCT/EP2010/070104
Authority: WO
Inventors: Anne Cathrine Bakken; Anita Sveen; Guro E. Lind; Ragnhild Lothe; Rolf L. Skotheim
Original assignee: Oslo Universitetssykehus Hf
Priority date: 2009-12-17
Filing date: 2010-12-17
Publication date: 2011-06-23

Abstract

The present inventors here present a novel strategy for identification of RNA transcript variants and demonstrate that these can be correlated to disease states in mammals such as cancer. In particular, the transcript variants show prevalence and specificity to cancer, and thus also show clinical applicability in e.g. cancer diagnostics and prognostics, treatment and therapeutics. The present inventors have identified RNA transcript variants of SLC39A14 that can be used as biomarkers. The RNA transcript variant may also be used as biomarkers for diagnosing, prognosing, monitoring, and or treatment selection for a cancer or the precursor to a cancer.

Description

TRANSCRIPT VARIANTS OF VNN1 AND SLC39A14

Technical field of the invention

The present invention relates to the identification of a new group of RNA transcript variants. In particular the present invention relates to RNA transcript variants comprising a 5' and/or 3' junction sequence(s) of a 5' outlier exon, wherein said junction sequence(s) comprises an intron or extra-genic originating expressed sequence. An object of the present invention relates to a method for the detection of an abnormal gene expression of at least one RNA transcript variant of SLC39A14. Another object of the present invention relates to the use of SLC39A14 RNA transcript variants as a biomarker. In particular the present invention relates to abnormal gene expressions in and biomarkers of cancer.

Background of the invention

Alternative splicing of primary transcripts (pre-mRNAs), alternative promoter usage, and alternative polyadenylation sites are mechanisms giving rise to multiple mRNA transcript variants and subsequently multiple protein isoforms per gene, and adding additional dimensions to the cellular complexity. Alterations of these normal processes are common in cancer cells and result in the production of mRNAs not existing in healthy cells or in the modification of tissue-specific ratios between normal mRNA types. One explanation for these differences is the fundamental difference in expression patterns of known splicing-regulatory genes in cancerous as compared to normal tissues. Individual cancer-specific variants may or may not be functionally important for the cells, but nevertheless, and due to the presence of sequences only present in malignant cells, they have the potential to function as therapeutic targets or as biomarkers for cancer diagnostics and prognostics. This great potential makes discovery and characterisation of novel transcript variants an interesting path towards a better understanding and management of cancer.

Different alternative splicing mechanisms are known. Exons which are either skipped or included in the final mRNA and are flanked by intron sequences on both sides are called cassette exons. Another mechanism of alternative splicing is the use of different 5' and 3' splice sites, where the amount of sequence included from a particular exon varies between different transcripts. If a splice site is missed by the splicing machinery, an intron can be retained in the final mRNA and contribute to the coding sequence. Also, some exons are mutually exclusive. This means that in the final processed mRNA, one out of two exons is always present, but never both.

In addition, different splice variants from the same gene may have completely different activities, because whole functional domains may be added or deleted from the protein-coding sequence. An example of such alterations is seen in the anti- apoptotic gene BIRC5. This gene is highly upregulated in various cancers and alternative splicing of its pre-mRNA produces four different mRNAs, which encode four different protein isoforms. One isoform has pro-apoptotic properties and acts like a naturally occurring antagonist of the anti-apoptotic functions of the other isoforms.

When discussing alternative core promoter usage it is important to keep in mind the differences between a transcription start site (TSS) and a core promoter. A gene's TSS is the first nucleotide to be transcribed into a particular RNA. The core promoter, on the other hand, is the genomic region that surrounds a TSS. The length of a core promoter is defined as the segment of DNA required to recruit the transcription initiation complex and initiate transcription, given the appropriate external signals. Alternative TSSs are often used within a core promoter.

Use of alternative core promoters enables diversification of transcriptional regulation within a single gene and thereby plays a significant role in the control of gene expression in various cell lineages, tissue types and developmental stages. The use of different core promoters can lead to two types of protein products, depending on the location of the translational start site relative to the used promoter. If the translational start site exists within the first exon, mRNA isoforms that encode distinct proteins will be produced. On the other hand, if the alternative first exon is non-coding, the alternative transcripts will have heterogeneous 5' untranslated regions (5'-UTR), which commonly implies different RNA stability, but the encoded proteins are identical. The molecular mechanisms behind the selective use of multiple promoters are not well known, but the use of diverse core promoter structures, variable concentrations of cis-regulatory elements and regional epigenetic mechanisms are thought to be important factors. Several oncogenes and tumour suppressor genes have multiple promoters and the aberrant use of one promoter over another in some of these genes is directly linked to cancerous cell growth.

The most common method for genome-wide gene expression analysis is by use of DNA microarrays. Here, the expression levels of genes are measured by

hybridisation signals to probes targeting predefined sequences. Thus, only exonic sequences known to the existing genome and transcriptome annotation are measured.

High-throughput sequencing of RNA (RNA-seq) is a powerful tool for identification of novel exons in individual samples. However, as of yet, a high cost make it unfeasible to process a large number of samples.

5' rapid amplification of cDNA ends (5'-RACE) is a method to detect transcript sequences 5' to a predefined gene-specific primer. In a large-scale effort to detect novel transcript structures, this method alone is in need of a good way to select candidate genes, the position of the RACE-primer, and the relevant samples to perform the RACE-experiments in.

Hence, an improved method for identification of novel RNA transcript variants would be advantageous, and in particular a more efficient and/or reliable method for identification of novel exons and exon-exon junction sequences in cancer samples would be advantageous.

The gene VNN1, encoding the vanin 1 protein, shares extensive sequence similarity with other members of the vanin gene family, which includes secreted and membrane-associated proteins. Detection of VNN1 expression was included in a blood-based biomarker panel for stratifying current risk for colorectal cancer (Marshall et al., Int. J. Cancer, 2009). The gene SLC39A14 encodes a protein belonging to a subfamily showing structural characteristics of zinc transporters. Two alternative exon 4 are known for this gene, 4A and 4B (Girijashanker et al., Mol. Pharmacol., 2008).

Summary of the invention Thus, an object of the present invention relates to a novel strategy for identification of transcript variants from a biological sample.

In particular, it is an object of the present invention to provide a method for identification of RNA transcript variants comprising a 5' and/or 3' junction

sequence(s) of a 5' outlier exon, wherein said junction sequence(s) comprises an intron or extra-genic originating expressed sequence in cancerous samples that solves the above mentioned problems of the prior art with regards to selection of candidate genes, selection of primer positions for RACE-PCR, and selection of the relevant samples with high likelihood of containing a novel transcript variant of the given candidate gene.

One aspect of the present invention relates to a method for the identification novel RNA transcript variant, by obtaining an exon expression profile of a gene in various test sample(s), obtaining a reference exon expression profile the gene in a reference sample, which may be taken from a control population such as a healthy population, identification of at least one 5' outlier exon, identification of 5' and/or 3' junction sequence(s) of said 5' outlier exon, and identification of RNA transcript variant comprising various parts of junction sequences.

Another aspect of the present invention relates to an RNA transcript variant comprising an 5' and/or 3' junction sequence(s) of an 5' outlier exon, wherein said junction sequence(s) comprises an intron or extra-genic originating expressed sequence.

Yet another aspect of the present invention relates to method for the detection of an abnormal gene expression pattern by identifying the novel RNA transcript variant comprising an 5' and/or 3' junction sequence(s) of an 5' outlier exon and comparing the expression level of such RNA transcript variant with a reference and correlating this to various diseases, such as cancer.

In conclusion, the present inventors here present a novel strategy for identification of these RNA transcript variants and furthermore demonstrate that these can be correlated to disease states in mammals. In particular the transcript variants show prevalence and specificity to cancer, and thus also show clinical applicability in e.g. cancer diagnostics, prognostics, treatment and therapeutics. In addition, the present invention relates to a method for the detection of abnormal gene expression of SLC39A14 RNA transcript variants, said method comprising identifying an expression level of at least one RNA transcript variant of SLC39A14 obtained from a test subject, comparing the expression level of said at least one RNA transcript variant of SLC39A14 with a reference obtained from a reference subject, selecting a desired sensitivity, selecting a desired specificity, and indicating the test subject as likely to have abnormal gene expression, if the expression level of the said at least one RNA transcript variant SLC39A14 in the sample obtained from a test subject is significantly different from the reference, and indicating the test subject as unlikely to have abnormal gene expression, if the expression level of said at least one RNA transcript variant of SLC39A14 is equal to the reference.

Another aspect of the present invention the abnormal expression pattern is indicative of cancer or a viral infection or a metabolic disease in the test subject.

Yet another aspect of the present invention relates to the use of at least one RNA transcript variant of SLC39A14 as a biomarker.

In an aspect of the present invention are these variants biomarkers for cancer or a precursor for cancer.

In an embodiment of the present invention is the cancer colorectal cancer or the precursor to cancer is colorectal adenomas.

Another aspect of the present invention relates to said biomarker as a biomarker for diagnosing, prognosing, and/or monitoring a cancer.

Brief description of the figures

Figure 1 :

Figure 1 shows representative nested RACE results from analysis of PRRX2,

RAD51L1, and VNN1. Lanes one, two, and three shows the results from nested RACE for PRRX2, RAD51L1, and VNN1, respectively. Abbreviations: Ml 500 base pair size marker; Nl, negative control for PRRX1; N2, negative control for

RAD51L1; N3, negative control for VNN1; M2, 100 base pair size marker.

Figure 2:

Figure 2 shows novel transcript variants of RAD51L1 in a colorectal cancer cell line. (A) Expression levels of the different probesets (often corresponding to the different exons) in RAD51L1 as seen from exon microarray data. Expression levels from the different cell lines are indicated by different shades and the thick lines represent the average for the six cell lines, ten colorectal carcinoma samples, and ten normal samples, respectively. The cell line SW48 deviates from the rest of the cell lines by showing stronger expression signals in the 3'-portion of the gene. (B) An overview of the different transcript variants. The black ruler on top indicates number of base pairs from the start of exon one. All exons are marked with a number. 1-14 indicates known exons or variants hereof, whereas α-η represents novel exons sequenced from SW48. The number of clones found with the same sequence is indicated in brackets after the name of the transcript. The start of every exon is in agreement with the number of base pairs from the start of exon one, but the exon width on the illustration is exaggerated for improved visualisation. Exons are numbered according to their location in the genomic sequence and exact positions of every exon can be found in Appendix II. Location of the nested gene-specific primer (NGSP) is shown by a black arrow. Five different transcripts are known according to Ensembl for RAD51L1. These transcripts have a total of 14 exons.

Figure 3:

Figure 3 shows results for NKAIN2. (A) Expression levels of the different exons in NKAIN2 for six cell lines. LS1034 has higher expression of exons eight to ten than the other cell lines. (B) Expression levels of the different exons in NKAIN2 for ten colorectal carcinomas. C1033III has higher expression of exons eight to ten than the other carcinomas. (C) An overview of the different transcript variants. Three different transcript variants are known for NKAIN2 according to Ensembl. Eight new transcripts were found by sequencing of the 5'-RACE products from LS1034 and C1033III and constitute a total of four new exons in introns four, eight, and nine. See legend of Figure 2 for more detailed explanations. Figure 4:

Figure 4 shows results for VNN1. (A) Expression levels of the different exons in VNN1 for six cell lines. HT29 deviates from the other cell lines by higher expression of exons six and seven. (B) An overview of the different transcript variants. One transcript with seven exons is known for VNN1. Three new transcript variants were found by sequencing of the 5'-RACE products from HT29 and include two new exons inside intron number five. See legend of Figure 2 for more detailed explanations.

Figure 5:

Figure 5 shows results for C4BPB. (A) Expression levels of the different exons in C4BPB for ten colorectal carcinoma samples. C1034III deviates from the rest in exons two to eight. (B) An overview of the different transcript variants. Five transcripts with a total of seven exons are known for C4BPB. Three new transcript variants were found by sequencing of the 5'-RACE products. See legend of Figure 2 for more detailed explanations.

Figure 6:

Figure 6 shows results for HOXCll. (A) Expression levels of the different exons in HOXCll for ten colorectal carcinoma samples. One sample, C1402III, deviates from the rest in the end of exon one and all of exon two. (B) An overview of the different transcript variants. One transcript with two exons is known for HOXCll. Two new transcript variants were found by sequencing of the 5'-end of the cDNA. See legend of Figure 2 for more detailed explanations.

Figure 7 :

Figure 7 shows results for TFR2. (A) Expression levels for the different exons in TFR2 for six cell lines. Two cell lines, SW48 and RKO, deviate from the rest in exons eight to eighteen. (B) An overview of the different transcript variants. One transcript with eighteen exons is known for TFR2. Ten new transcript variants were found by sequencing of the 5'-end of the cDNA. See legend of Figure 2 for more detailed explanations.

Figure 8:

Figure 8 shows results for SERPINB7. (A) Expression levels of the different exons in SERPINB7 for six cell lines. One cell line, LS1034, deviates from the rest in exons five to nine. (B) An overview of the different transcript variants. Two transcripts with a total of nine exons are known for SERPINB7. Three transcript variants were found by sequencing of the 5'-RACE products in LS1034. See legend of Figure 2 for more detailed explanations.

Figure 9:

Figure 9 shows results for TFPT. (A) Expression levels of the different exons in TFPT for six cell lines. One cell line, SW48, deviates from the rest in exons four to seven. (B) An overview of the different transcript variants. Four different transcripts with seven exons are known for TFPT. Two transcript variants were found by sequencing of the 5'-RACE products from SW48. See legend of Figure 2 for more detailed explanations.

Figure 10:

Figure 10 shows results for GJB6. (A) Expression levels of the different exons in GJB6 for six cell lines. One cell lines, HT29, deviates from the others by higher expression of exons five and six. (B) An overview of the different transcript variants. Four different transcripts with a total of six exons are known for GJB6. Six transcript variants were found by sequencing of the 5'-RACE products from HT29. See legend of Figure 2 for more detailed explanations.

Figure 11 :

Figure 11 shows results for PRRX1. (A) Expression levels of the different exons in PRRX1 for six cell lines. One cell line, SW48, deviates from the others by higher expression of exons two to five. (B) Overview of the different transcript variants. Two different transcripts with a total of five exons are known for PRRX1. Eight transcript variants were found by sequencing of the 5'-RACE products from SW48. See legend of Figure 2 for more detailed explanations.

Figure 12:

Figure 12 shows results for PRRX2. (A) Expression levels of the different exons in PRRX2 for ten colorectal carcinoma samples. One sample, C1033III, deviates from the others by higher expression of exon number four. (B) An overview of the different transcript variants. One transcript with four exons is known for PRRX2 and two transcript variants were found by sequencing of the 5'-RACE products from C1033III. See legend of Figure 2 for more detailed explanations.

Figure 13:

According to the latest version of Ensembl (release 56, September 2009), there is only one transcript annotated for VNNl. This variant, ENST00000367928, has seven exons. Three new transcript variants were found by sequencing of the 5'-RACE products from HT29 and include two new exons inside intron number five of ENST00000367928. To distinguish transcripts including the novel exons from the ENST00000367928 transcripts, an RT-PCR assay was developed with two specific forward primers and a common reverse primer. The forward primer targeting ENST00000367928 is specific to the annotated exon 5 and the forward primer targeting the novel transcripts target the common region in exon {alpha}.

Figure 14:

RT-PCR of VNNl with primers specifically binding to the ENST00000367928 exon 5 to 6 from 8 normal colon mucosa (marked "N"), 105 colorectal cancers, 2 negative controls (marked "Neg"). PCR-product at the expected length was detected for all samples.

Figure 15:

RT-PCR of novel exons within VNNl with one of the primers specifically binding to the novel exon {alpha} and the exon 6 in ENST00000367928 from 8 normal colon mucosa (marked "N"), 105 colorectal cancers, 2 negative controls (marked "Neg"). PCR-products at the expected lengths according to the 5'RACE experiments, and one additional band, were detected for 87 % of the colorectal cancers, but not for any of the normal colon mucosa or the negative controls.

Figure 16:

Expression levels of the different probe selection regions (similar to individual exons) for SLC39A14 as seen from exon microarray data. The light gray and dark gray lines represent the log-2 averages of the normal and cancer tissue samples, respectively. Exons are numbered according to ENST00000289952; however, exon four-primed is not present in this transcript, but in another transcript of this gene. (B) Two known splicing events assumed to be responsible for the interesting exon- wise plot are depicted. The light gray and dark gray lines represent the splicing events dominating in normal and cancer tissues, respectively. The two mutually exclusive exons four have identical size and similar sequences. Two real-time RT- PCR assays were designed with identical primers but distinct probes, as depicted.

Figure 17 :

First-line validation of SLC39A14 transcript variants by real-time RT-PCR. To the left, the real-time RT-PCR results (amplification plots) obtained with the assay containing the probe annealing to exon four-primed, which from the exon-wise plot seems to have a higher level of inclusion in normal tissue, are shown for all the samples included in the first-line validation; three normals, 16 CRC and six colon cancer cell lines. To the right, the corresponding data obtained with the probe hybridising to exon four are depicted. The fluorescence intensities (y-axis) are plotted against the number of PCR cycles (x-axis). The red horizontal line indicates the Cycle threshold (Ct).

Figure 18:

SLC39A14 real-time RT-PCR data in a series of clinical CRC and normal tissue samples, in addition to six colon sell lines, showing the relative expression between the two mutually exclusive splicing variants. The Ct values obtained in the assay with a probe in exon four-primed ("normal exon") is normalised against the Ct values obtained in the assay with a probe in exon four ("cancer exon"), and this gives the relative expression along the y-axis (log-2) for each of the samples (x- axis). It is worth mentioning that all samples that did not cross the Ct line before cycle 34 were considered not to express the specific transcript, and were given the value 34 prior to the calculations.

Figure 19:

Expression levels of the different probe selection regions (similar to individual exons) for SLC39A14 as seen from exon microarray data. The bright gray and dark gray lines represent the log-2 averages of the normal colonic mucosa and colorectal cancer tissue samples, respectively. When comparing these expression averages to each other, the exon 4A has a higher relative expression average in normal colonic mucosa, whereas the exon 4B has a higher relative expression average in the colorectal cancer. Exons are numbered according to Ensembl transcripts ENST00000359741 and ENST00000381237 (Ensembl release 60 - Nov 2010).

These are identical transcripts, with the exception of the fourth exon, which are different in the two. Here, we label the exon four in ENST00000359741 as exon 4A and the exon four in ENST00000381237 as exon 4B. Exon 4A has the exon identifier ENSE00001401146 and exon 4B has the identifier ENSE00000683833. (B) Two known splicing events assumed to be responsible for the interesting exon-wise plot are depicted. The bright gray and dark gray lines represent the splicing events dominating in normal colonic mucosa and colorectal cancer tissues, respectively. The two mutually exclusive exons four, 4A and 4B, have identical sizes and similar, but not identical, sequences. Two real-time RT-PCR assays were designed with identical primers but distinct probes, as depicted.

Figure 20:

First-line validation of SLC39A14 transcript variants by TaqMan real-time RT-PCR. To the left, the real-time RT-PCR results (amplification plots) obtained with the assay containing the probe annealing to exon 4A, which from the exon-wise plot seems to have a higher level of inclusion in normal colonic mucosa than in colorectal cancers, are shown for all the samples included in the first-line validation; three normals, 16 CRC and six colon cancer cell lines. To the right, the

corresponding data obtained with the probe hybridising to exon 4B are depicted. The fluorescence intensities (y-axis) are plotted against the number of PCR cycles (x-axis). The red horizontal line indicates the Cycle threshold (Ct).

Figure 21 :

SLC39A14 real-time RT-PCR data from healthy and disease samples from both colorectal and a variety of additional sites, showing the relative expression between the two mutually exclusive splicing variants. The Ct values obtained in the assay with a probe in exon 4B is normalised against the Ct values obtained in the assay with a probe in exon 4A, and this gives the relative expression along the y-axis (log-2) for each of the samples (x-axis). It is worth mentioning that all samples that did not cross the Ct line before cycle 34 were considered not to express the specific transcript, and were given the value 34 prior to the calculations. Figure 22:

RNA-sequencing data quantifying expression levels from exons 4A and 4B of SLC39A14. The samples are from left to right, six colorectal cancer (CRC) cell lines, two CRC tissue samples, their two matched normal colonic mucosa, a healthy lymph node, and healthy white blood cells. The method used is paired-end RNA- sequencing by the Solexa technology of Illumina, and processed by the Genome Analyzer IIx machine.

The present invention is described in more detail in the following. Detailed description of the invention

The present invention provides methodology, which is employed in a screening strategy for the identification of transcript variants from a biological sample. The strategy includes the following objectives:

• Investigate the expression level of individual exons in candidate genes, or in all genes in the genome, as from automated analyses of exon microarray data.

• Investigate the 5'-end of mRNA from individual genes in cell lines and/or tumour samples where exon expression profile in the 3'-end of the gene that is different from that of a reference profile.

The strategy was established for candidate gene selection of genes with outlier expression profiles in colorectal cancer and for genes with known or putative involvement as oncogenic fusion transcripts. For all candidate genes, the expression levels across all exons were investigated, and genes with overexpression selectively from 3'-exons were further analysed for novel upstream sequences.

In total, the exon expression levels for 508 genes were investigated. Eleven of these genes had deviating exon expression profiles indicating qualitative changes in the transcript structure and were therefore further investigated. RNA transcript variants were identified in all of the eleven genes. These included potentially new promoters, novel exons within intron sequences and intron retentions, however, no fusion genes were found. In conclusion, the present inventors here present methods for identification of RNA transcript variants and furthermore demonstrate that these can be correlated to disease states in mammals. In particular the transcript variants show prevalence and specificity to cancer, and thus also show clinical applicability in e.g. cancer diagnostics, prognostics, treatment and therapeutics.

Thus, one aspect of the present invention relates to a method for the identification of at least one RNA transcript variant, said method comprising obtaining an exon expression profile of a gene of interest in a test sample, obtaining a reference exon expression profile of said gene in a reference sample, identification of at least one 5' outlier exon, identification of 5' and/or 3' junction sequence(s) of said 5' outlier exon, and identification of at least one RNA transcript variant comprising at least one of said junction sequences.

Exon expression profile

The exon expression profile as used herein refers to the individual expression measurements from two or more exons along a gene of interest. The expression profiles represent the abundance of the individual exons in the pool of RNA transcripts present in a sample. The expression measurements are reported as relative expression as compared to the corresponding exon expression profile of a reference. Such an exon expression profile is obtained from RNA or single/double- stranded cDNA. The profile can be obtained as an average expression from 1 to ~ n number of samples.

Analysis of the exon expression profile turned to be an important step in the process of enriching for genes with alterations in their transcript structures. If every exon in a gene is under the control of the same promoter, it is expected that the exon expression levels to be similar throughout the gene.

If, on the other hand, a gene has a second alternative promoter, the exons downstream of the new promoter/breakpoint will be under the control of a different promoter than the upstream exons. The 5'-portion of the original gene is therefore regulated by one promoter and the 3'-portion by another, leading to different expression of the two parts. This may give rise to longitudinal exon expression profiles looking like the ones seen in Figure 2A to Figure 12A, where exons in the 3'-end of a gene have higher expression than the 5'-exons in certain samples as compared to others.

Thus, an expression profile of a sample as compared to that of a reference can be compared statistically.

The statistical significance may be determined by the standard statistical methodology known by the person skilled in the art.

Outlier transcript profile

An outlier transcript profile refers to a transcript profile, where the relative exon expression profile of the test sample vs. the reference sample is higher in the 3'- portion of the transcript (one or more exons at the 3'-end) as compared to the 5'- end of the transcript (one or more exons at the 5'-end) with statistical significance.

An embodiment of the present invention refers to a an outliner transcript profile, wherein the relative profile of the test sample vs. the reference sample is significantly higher in the 3'-portion of the transcript (one or more exons at the 3'- end) as compared to the 5'-end of the transcript (one or more exons at the 5'-end) with a confidence interval of 50%, such as 75%, such as 90%, such as 95%, such as 99%.

The significance may be determined by the standard statistical methodology known by the person skilled in the art.

Identification of at least one 5' outlier exon

Identification of at least one 5' outlier exon as used herein refers to the

identification of at least the first 5' outlier exon in an exon expression profile. One method for identification of an exon expression profile indicating the existence of such 5' exon can be through calculation of two probabilities for each exon-exon junction. A first probability is based on a t-test for whether values from all upstream and all downstream exons are likely to belong to different populations

[P(transcript)]. A second probability is based on a t-test for whether the values from the immediate up- and downstream exons are likely to belong to different populations [P(exon)]. A Transcript breakpoint score (TBS) is calculated as the product of the two [TBS = P(transcript) * P(exon)].

In statistics is a confidence interval (CI) or confidence bound is an interval estimate of a population parameter. Instead of estimating the parameter by a single value, an interval likely to include the parameter is given. Thus, confidence intervals are used to indicate the reliability of an estimate. How likely the interval is to contain the parameter is determined by the confidence level or confidence coefficient.

Increasing the desired confidence level will widen the confidence interval. For example, a CI can be used to describe how reliable survey results are. A 95% confidence interval for the proportion in the whole population having the same intention on the survey date might be 36% to 44%. All other things being equal, a survey result with a small CI is more reliable than a result with a large CI and one of the main things controlling this width in the case of population surveys is the size of the sample questioned. Confidence intervals and interval estimates more generally have applications across the whole range of quantitative studies.

If a statistic is presented with a confidence interval, and is claimed to be statistically significant, the underlying test leading to that claim will have been performed at a significance level of 100% minus the confidence level of the interval.

Accordingly an embodiment of the present invention refers to a method for identification of a 5' outlier exon of the invention that can be indentified through calculation of two probabilities for each exon-exon junction. One probability is based on a t-test for whether values from all upstream and all downstream exons are likely to belong to different populations [P(transcript)]. A second probability is based on a t-test for whether the values from the immediate up- and downstream exons are likely to belong to different populations [P(exon)]. A Transcript breakpoint score (TBS) is calculated as the product of the two [TBS = P(transcript) * P(exon)] with a confidence interval of 50%, such as 75%, such as 90%, such as 95%, such as 99%.

Intron or extra-genic originating expressed transcripts

Intron or extra-genic originating expressed sequence also referred to as intergenic sequences as used herein refers to novel transcript sequences that have previously been annotated as intronic or intergenic or a sequence that have not been annotated before. That is, Ensembl and RefSeq do not consider these sequences as part of the reference transcripts of the human genome.

Expressed transcript

An expressed transcript as used herein refers to a transcript that is encoded by a gene and expressed to form a transcript RNA. This RNA can be coding, or non- coding.

Junction sequence

A junction according to the present invention refers to the intersection of genetic elements such as exons and introns. Accordingly, the junction sequence refers to the sequence spanning the flanking sequence of the junction. Thus, the junction sequence of two juxtaposing exons in a mRNA comprises the 3' flanking sequence of the 5' exon and the 5' flanking sequence of the 3' exon.

Hence, the 5' junction sequence of a particular exon will contain at least part of the 5' end of the exon of interest and at least part of the 3' flanking sequence of the 5' exon. Similarly will the 3' junction sequence of an exon contain at least part of the 3' end of the exon of interest and at least part of 5' flanking sequence of the 5' exon.

In an embodiment 5' and/or the 3' junction sequences of the present invention are identified by sequencing of a polynucleotide obtained from RACE, one-sided PCR and/or anchored PCR.

In one embodiment the 5' flanking sequence is less than 15kb, such as less than lOkb, for example less than such as lOkb, for example less than such as 5 kb, for example less than such as 4kb, for example less than such as 3kb, for example less than such as 2kb, for example less than such as lkb, for example less than such as 500b.

In one embodiment the 3' flanking sequence is less than 15kb, such as less than lOkb, for example less than such as lOkb, for example less than such as 5 kb, for example less than such as 4kb, for example less than such as 3kb, for example less than such as 2kb, for example less than such as lkb, for example less than such as 500b. RNA transcript variant

An aspect of the present invention relates to an RNA transcript variant comprising an 5' and/or 3' junction sequence(s) of an 5' outlier exon, wherein said junction sequence(s) comprises an intron or extra-genic originating expressed sequence.

Another aspect of the present invention relates to an isolated RNA transcript variant obtained from a method for the identification of at least one RNA transcript variant, said method comprising obtaining an exon expression profile of a gene of interest in a test sample, obtaining a reference exon expression profile of said gene in a reference sample, identification of at least one 5' outlier exon, identification of 5' and/or 3' junction sequence(s) of said 5' outlier exon, and identification of at least one RNA transcript variant comprising at least one of said junction sequences.

A transcription start site TSS of a gene is the first nucleotide to be transcribed into a particular RNA. The core promoter, on the other hand, is the genomic region that surrounds a TSS. The length of a core promoter is defined as the segment of DNA required to recruit the transcription initiation complex and initiate transcription, given the appropriate external signals. Alternative TSSs are often used within a core promoter. Thus, the RNA transcripts, which are products of transcriptional initiation from different TTSs, will have different terminal 5' flanking sequences.

In one embodiment the RNA transcript variant is the transcriptional product of a core promoter. The core promoter may be activated by various stimuli and the aberrant core promoter activity may correlate with clinical conditions such as cancer, viral infections and metabolic conditions.

A 5' cap structure is found on the 5' end of an mRNA molecule and consists of a 7- methylguanosine connected to the mRNA via a 5' to 5' triphosphate linkage.

In one embodiment the junction is the 5' to 5' triphosphate bridge linking the 7- methylguanosine to 5' end of the RNA transcript variant. Accordingly, in one particular embodiment the junction sequences is the 5' flanking sequences of the 5' outlier exon and 7-methylguanosine linked by the 5' to 5' triphosphate bridge. This structure is the 5' capture and the 5' terminal sequences of the 5' outlier exon, which identifies the RNA transcript variant of the embodiment. RNA transcript variant as used herein refers to any RNAs that comprises exons, introns or part hereof originating from the same gene. The RNA transcript variant can arise through alternative or aberrant pre-mRNA processing, alternative or aberrant promoter usage or polyadenylation initiation sites.

This means that one or more exons or exon-junctions are differentially included in the RNA transcript variants of a particular gene. Thus, can the RNA transcript variants be one exon, two exons, three exons, or more exons of a particular gene.

RNA transcript variants can result in polypeptides, but can also be non-coding. Expression level

The expression level of a given genetic element as used herein refers to the absolute or relative amount of RNA corresponding to this genetic element in a given sample. Expressed genes include genes that are transcribed into mRNA and then translated into protein, as well as genes that are transcribed into mRNA, or other types of RNA such as, tRNA, rRNA or other non-coding RNAs, that are not translated into protein. RNA expression is a highly specific process which can be monitored by detecting the absolute or relative RNA levels.

Thus, the expression level refers to the amount of RNA in a sample. The expression level is usually detected using microarrays, northern blotting, RT-PCR, SAGE, RNA- seq, or similar RNA detection methods.

When expression levels of a specific RNA in a test sample is compared to a reference sample they can either be different or equal. However, using today's detection techniques is an exact definition of different or equal result can be difficult because of noise and variations in obtained expression levels from different samples. Hence, the usual method for evaluating whether two or more expression levels are different or equal involves statistics.

Statistics enables evaluation of significantly different expression levels and significantly equal expressions levels. Statistical methods involve applying a function/statistical algorithm to a set of data. Statistical theory defines a statistic as a function of a sample where the function itself is independent of the sample's distribution : the term is used both for the function and for the value of the function on a given sample. Commonly used statistical tests or methods applied to a data set include t-test, f-test or even more advanced test and methods of comparing data. Using such a test or methods enables a conclusion of whether two or more samples are significantly different or significantly equal.

Abnormal gene expression pattern

The expression of a gene results in at least one RNA transcript. As used herein an abnormal gene expression pattern refers to a significantly different expression level of a gene in a test sample as compared to a reference sample.

An embodiment of the present invention refers to an abnormal gene expression pattern refers to a significantly different expression level of a gene in a test sample as compared to a reference sample with a confidence interval of 50%, such as 75%, such as 90%, such as 95%, such as 99%.

Accordingly, one embodiment relates to a method for the identification of at least one RNA transcript variant, wherein the expression of the 5' outlier exon is significantly higher than the corresponding 5' exon of the reference.

Accordingly, one embodiment relates to a method for the identification of at least one RNA transcript variant, wherein the expression of the 5' outlier exon is significantly lower than the corresponding 5' exon of the reference.

In a further embodiment the expression level of each of the 3' exons from said test sample are higher than their corresponding 3' exons of the reference.

Correlation of abnormal gene expression to a disease state

Another aspect of the invention relates to method for the detection of an abnormal gene expression pattern, said method comprising identifying an expression level of an RNA transcript variant comprising an 5' and/or 3' junction sequence(s) of an 5' outlier exon, wherein said junction sequence(s) comprises an intron or extra-genic originating expressed sequence in a sample obtained from a test subject, comparing the expression level of said RNA transcript variant with a reference obtained from a reference subject, selecting a desired sensitivity, selecting a desired specificity, and indicating the test subject as likely to have an abnormal gene expression pattern, if the expression level of the RNA transcript variant in the sample obtained from a test subject is significantly different from the reference, and indicating the test subject as unlikely to have an abnormal gene expression pattern, if the expression level of the RNA transcript variant is equal to the reference.

Another aspect of the present invention relates to a method for the detection of an abnormal gene expression of at least one gene, wherein said at least one gene is selected from the group consisting of VNN1 and SLC39A14, said method comprising identifying an expression level of at least one RNA transcript variant of said at least one gene in a sample obtained from a test subject, comparing the expression level of said at least one RNA transcript variant of said at least one gene with a reference obtained from a reference subject, selecting a desired sensitivity, selecting a desired specificity, indicating the test subject as likely to have abnormal gene expression, if the expression level of said at least one RNA transcript variant of said at least one gene in the sample obtained from a test subject is significantly different from the reference, and indicating the test subject as unlikely to have abnormal gene expression, if the expression level of said at least one RNA transcript variant of said at least one gene is equal to the reference.

In an embodiment relates to the method for the detection of an abnormal gene expression of at least one gene, such as one gene, such as two genes, such as three genes, such as four genes, such as five genes.

Yet another aspect of the present invention relates to a method for the detection of abnormal gene expression of at least one gene, wherein said at least one gene is selected from the group consisting of VNN1 and SLC39A14, said method comprising the step of determining an expression level of at least one RNA transcript variant of said at least one gene in a sample obtained from a test subject.

Another aspect of the present invention relates to a method for the detection of abnormal gene expression of at least one gene, wherein said at least one gene is selected from the group consisting of VNN1 and SLC39A14, said method comprising the step of determining an expression level of at least one RNA transcript variant of said at least one gene in a sample obtained from a test subject further comprising the steps of comparing the expression level of said at least one RNA transcript variant of said at least one gene with a reference obtained from a reference subject, selecting a desired sensitivity, selecting a desired specificity, indicating the test subject as likely to have an abnormal gene expression, if the expression level of said at least one RNA transcript variant of said at least one gene in the sample obtained from a test subject is significantly different from the reference, and indicating the test subject as unlikely to have an abnormal gene expression, if the expression level of said at least one RNA transcript variant of said at least one gene is equal to the reference.

In an embodiment of the present invention is the expression level of the at least one RNA transcript variant of the gene in the test subject higher than the reference subject.

In another embodiment of the present invention is the RNA transcript variant selected from the group consisting of VNN1 A (SEQ ID NO: 15), VNN1 B (SEQ ID NO: 16), and VNN1 C (SEQ ID NO: 17), and the SLC39A14 transcript variant is selected from the group consisting of transcript 1 (SEQ ID NO: 137), transcript 2 (SEQ ID NO: 138), or transcript 3 (SEQ ID NO: 139).

In another embodiment of the present invention comprises the RNA transcript variant one or more of the exons selected from the group consisting of VNNla (SEQ ID NO: 131), VNNla' (SEQ ID NO: 132), VNNla" (SEQ ID N0133), νΝΝΙβ (SEQ ID NO: 134), and νΝΝΙβ' (SEQ ID NO: 135), and 4 is ENSE00000683833 (SEQ ID NO: 136), and 4A is ENSE00001401146 (SEQ ID NO: 144).

The appearance or increase of a RNA transcript variant in a cell such as a neoplastic cell for example a tumour cell indicates a phenotypic change of the cells present in a sample obtained from said subject compared to a the corresponding cells in a sample from a reference subject.

The appearance or increase of a RNA transcript variant in cells of a sample obtained from neoplastic tissue for example a tumour tissue may therefore be indicative of a gain-of-function of an oncogene involved in the progression of carcinogenesis of the tumour. Accordingly, the RNA transcript variant is a potential candidate biomarker applicable for the diagnosis of the diseased state i.e. cancer.

In additionally embodiment can the RNA transcript variant be used as a biomarker for the progression of the disease state by monitoring of differential expression patterns over time.

Accordingly will the RNA transcript variant be applicable for diagnosis, prognosis and a treatment of clinical conditions or a diseased state.

Thus, in an embodiment is the expression level of an RNA transcript variant in the test subject is significantly higher or lower than the reference subject.

In another embodiment is the expression level of each of the 3' exons from said test sample higher than their corresponding 3' exons of the reference.

In an embodiment the expression level of an RNA transcript variant is applicable for the diagnosis of a diseased state i.e. cancer, a viral infection or a metabolic disease in the test subject.

In another embodiment of the present invention, the abnormal expression pattern is indicative of cancer or an inflammatory disease or a viral infection or a metabolic disease in the test subject.

In a specific embodiment of the present invention, the cancer is selected from the group consisting of colorectal cancer, prostate cancer, breast cancer, lung cancer, liver cancer, kidney cancer, ovarian cancer, endometrial cancer, pancreatic cancer, brain cancer, testicular cancer, leukemia, lymphoma, sarcoma.

In a specific embodiment of the present invention is the cancer is colorectal cancer or the precursor to cancer is colorectal adenomas.

Colorectal cancer includes cancerous growths in the colon, rectum and appendix. Colorectal cancers arise from adenomatous polyps in the colon. Adenomatous polyps are usually benign, but some develop into cancer over time. Early

identification of the adenomatous polyps that have the potential to develop into cancer is therefore of great importance.

The identification of dysplastic cells or polyps from other inflammatory conditions like inflammatory bowel disease (IBD) and Crohn's disease is difficult and is usually done by morphological evaluation by a pathologist.

Localized colon cancer is usually diagnosed through colonoscopy, and invasive cancers that are confined within the wall of the colon (TNM

(tumors/nodes/metastases) stages I and II) are curable with surgery.

If untreated, they spread to regional lymph nodes (stage III), where some are curable by surgery and chemotherapy. Cancer that metastasizes to distant sites (stage IV) is usually not curable.

An aspect of the present invention is related to the identification of dysplastic cells or adenomatous polyps that are likely to develop into cancer.

In an embodiment of the present invention is the expression of SLC39A14 transcript variants 1, 2 and 3 used in the identification of adenomatous polyps or dysplastic cells that are likely to develop into cancer.

Another aspect of the present invention relates to the use of SLC39A14 transcript variants 1, 2 and 3 in the identification of adenomatous polyps or dysplastic cells that are likely to develop into cancer.

In an embodiment of the present invention are the adenomatous polyps or dysplastic cells that are likely to develop into cancer identified in a subject that is suffering from an inflammatory state of the colorectal region.

Such inflammatory state can be inflammatory bowel disease (IBD) like ulcerative colitis (UC) and Crohn's disease. One aspect of the present invention relates to a method of the present invention, wherein the 4B exon is present in the sample or test material.

In an embodiment of the present invention is the 4A exon is not present in the sample or test material.

Yet another aspect of the present invention relates to a method for identification of an abnormal expression pattern of SLC39A14 which is indicative of a precursor of colorectal cancer.

In an embodiment of the present invention is the likelihood of development into cancer evaluated by correlating an abnormal SLC39A14 expression pattern to a diseased state.

In an embodiment of the present invention is the expression of SLC39A14 exon 4B or the transcript variants 1, 2, and 3 as such, used for early detection of colorectal cancer or precursor lesions of colorectal cancer.

The test material may for example be a peripheral blood sample, stool sample, or a bowel biopsy.

In an embodiment of the present invention is the expression of SLC39A14 exon 4B or the transcript variants 1, 2, and 3 as such, used in the monitoring of disease after treatment of colorectal cancer i.e. testing for remnants of cancer cells and/or relapse.

This test material may for example be a peripheral blood sample, stool sample, or a bowel biopsy.

In an embodiment of the present invention is the expression of SLC39A14 exon 4B or the transcript variants 1, 2, and 3 as such, used for improved staging of colorectal cancer.

RNA or protein measurements, including, but not limited to, RNA in situ

hybridization, qRT-PCR, and immunohistochemistry, are used to test for the presence of cancer cells, indicative of stage III cancer. In an aspect of the present invention relates to the genomic genes that incode the RNA transcript variants of the present invention. The RNA transcript variants can be detected in the genomic DNA using standard DNA assaying techniques that are known in the art.

Thus relates one aspect of the present invention to detection and/or correlation of the genomic DNA encoding the RNA transcript variants of the present invention with cancer or an inflammatory disease or a viral infection or a metabolic disease in the test subject.

One embodiment of the present invention relates to an isolated nucleic acid molecule selected from the group consisting of SEQ ID NO 1, SEQ ID NO 2, SEQ ID NO 3, SEQ ID NO 4, SEQ ID NO 5, SEQ ID NO 6, SEQ ID NO 7, SEQ ID NO 8, SEQ ID NO 9, SEQ ID NO 10, SEQ ID NO 11, SEQ ID NO 12, SEQ ID NO 13, SEQ ID NO 14, SEQ ID NO 15, SEQ ID NO 16, SEQ ID NO 17, SEQ ID NO 18, SEQ ID NO 19, SEQ ID NO 20, SEQ ID NO 21, SEQ ID NO 22, SEQ ID NO 23, SEQ ID NO 24, SEQ ID NO 25, SEQ ID NO 26, SEQ ID NO 27, SEQ ID NO 28, SEQ ID NO 29, SEQ ID NO 30, SEQ ID NO 31, SEQ ID NO 32, SEQ ID NO 33, SEQ ID NO 34, SEQ ID NO 35, SEQ ID NO 36, SEQ ID NO 37, SEQ ID NO 38, SEQ ID NO 39, SEQ ID NO 40, SEQ ID NO 41, SEQ ID NO 42, SEQ ID NO 43, SEQ ID NO 44, SEQ ID NO 45, SEQ ID NO 46, SEQ ID NO 47, SEQ ID NO 48, SEQ ID NO 49, SEQ ID NO 50, SEQ ID NO 51, SEQ ID NO 52, SEQ ID NO 53, SEQ ID NO 54, SEQ ID NO 131, SEQ ID NO 132, SEQ ID NO 133, SEQ ID NO 134, SEQ ID NO 135, SEQ ID NO 136, SEQ ID NO 137, SEQ ID NO 138, SEQ ID NO 139, that can be correlated to an abnormal gene expression pattern.

The above sequences are identified using the methodology of the present invention described herein. Thus, these sequences represent RNA transcript variants that are present and/or expressed to a higher level than the reference sample.

Accordingly, an embodiment of the invention relates to a biomarker selected from the group consisting of SEQ ID NO 1, SEQ ID NO 2, SEQ ID NO 3, SEQ ID NO 4, SEQ ID NO 5, SEQ ID NO 6, SEQ ID NO 7, SEQ ID NO 8, SEQ ID NO 9, SEQ ID NO 10, SEQ ID NO 11, SEQ ID NO 12, SEQ ID NO 13, SEQ ID NO 14, SEQ ID NO 15, SEQ ID NO 16, SEQ ID NO 17, SEQ ID NO 18, SEQ ID NO 19, SEQ ID NO 20, SEQ ID NO 21, SEQ ID NO 22, SEQ ID NO 23, SEQ ID NO 24, SEQ ID NO 25, SEQ ID NO 26, SEQ ID NO 27, SEQ ID NO 28, SEQ ID NO 29, SEQ ID NO 30, SEQ ID NO 31, SEQ ID NO 32, SEQ ID NO 33, SEQ ID NO 34, SEQ ID NO 35, SEQ ID NO 36, SEQ ID NO 37, SEQ ID NO 38, SEQ ID NO 39, SEQ ID NO 40, SEQ ID NO 41, SEQ ID NO 42, SEQ ID NO 43, SEQ ID NO 44, SEQ ID NO 45, SEQ ID NO 46, SEQ ID NO 47, SEQ ID NO 48, SEQ ID NO 49, SEQ ID NO 50, SEQ ID NO 51, SEQ ID NO 52, SEQ ID NO 53, SEQ ID NO 54, SEQ ID NO 131, SEQ ID NO 132, SEQ ID NO 133, SEQ ID NO 134, SEQ ID NO 135, SEQ ID NO 136, SEQ ID NO 137, SEQ ID NO 138, SEQ ID NO 139, SEQ ID NO 144.

A biomarker can be a marker for a diseased state i.e. cancer, a viral infection, a metabolic disease or an inflammatory disease in the test subject.

In another embodiment of the present invention, the biomarker is indicative of cancer or a viral infection or a metabolic disease in the test subject.

In a specific embodiment of the present invention, the cancer is selected from group consisting of colorectal cancer, prostate cancer, breast cancer, lung cancer, liver cancer, kidney cancer, ovarian cancer, endometrial cancer, pancreatic cancer, brain cancer, testicular cancer, leukemia, lymphoma, sarcoma.

An aspect of the present invention relates to the use of at least one RNA transcript variant selected from the list consisting of (SEQ ID NO: 15), (SEQ ID NO: 16), (SEQ ID NO: 17), (SEQ ID NO: 18), (SEQ ID NO: 131), (SEQ ID NO: 132), (SEQ ID

NO: 133), (SEQ ID NO: 134), (SEQ ID NO: 135), (SEQ ID NO: 136), (SEQ ID NO: 137) (SEQ ID NO: 138), (SEQ ID NO: 139), and (SEQ ID NO: 144) as a biomarker.

Another aspect of the present invention relates to the use of the biomarker as a biomarker for diagnosing, prognosing, and/or monitoring a cancer.

Another aspect of the present invention relates to the use of the biomarker as a biomarker for diagnosing, prognosing, and/or monitoring a cancer, wherein the cancer is selected from group consisting of colorectal cancer, prostate cancer, breast cancer, lung cancer, liver cancer, kidney cancer, ovarian cancer, endometrial cancer, pancreatic cancer, brain cancer, testicular cancer, leukemia, lymphoma, sarcoma. In an aspect of the present invention is the biomarker used for identification of adenomatous polyps or dysplastic cells that are likely to develop into cancer.

A further embodiment of the invention relates to an isolated nucleic acid molecule selected from the group consisting of SEQ ID NO 1, SEQ ID NO 2, SEQ ID NO 3, SEQ ID NO 4, SEQ ID NO 5, SEQ ID NO 6, SEQ ID NO 7, SEQ ID NO 8, SEQ ID NO 9, SEQ ID NO 10, SEQ ID NO 11, SEQ ID NO 12, SEQ ID NO 13, SEQ ID NO 14, SEQ ID NO 15, SEQ ID NO 16, SEQ ID NO 17, SEQ ID NO 18, SEQ ID NO 19, SEQ ID NO 20, SEQ ID NO 21, SEQ ID NO 22, SEQ ID NO 23, SEQ ID NO 24, SEQ ID NO 25, SEQ ID NO 26, SEQ ID NO 27, SEQ ID NO 28, SEQ ID NO 29, SEQ ID NO 30, SEQ ID NO 31, SEQ ID NO 32, SEQ ID NO 33, SEQ ID NO 34, SEQ ID NO 35, SEQ ID NO 36, SEQ ID NO 37, SEQ ID NO 38, SEQ ID NO 39, SEQ ID NO 40, SEQ ID NO 41, SEQ ID NO 42, SEQ ID NO 43, SEQ ID NO 44, SEQ ID NO 45, SEQ ID NO 46, SEQ ID NO 47, SEQ ID NO 48, SEQ ID NO 49, SEQ ID NO 50, SEQ ID NO 51, SEQ ID NO 52, SEQ ID NO 53, SEQ ID NO 54, SEQ ID NO 131, SEQ ID NO 132, SEQ ID NO 133, SEQ ID NO 134, SEQ ID NO 135, SEQ ID NO 136, SEQ ID NO 137, SEQ ID NO 138, SEQ ID NO 139, SEQ ID NO 144 which encodes a polypeptide.

An embodiment of the present invention relates to antibodies raised against the polypeptides of the present invention and use hereof for therapeutic purposes.

A further embodiment the invention relates to an isolated nucleic acid molecule selected from the group consisting of SEQ ID NO 1, SEQ ID NO 2, SEQ ID NO 3, SEQ ID NO 4, SEQ ID NO 5, SEQ ID NO 6, SEQ ID NO 7, SEQ ID NO 8, SEQ ID NO 9, SEQ ID NO 10, SEQ ID NO 11, SEQ ID NO 12, SEQ ID NO 13, SEQ ID NO 14, SEQ ID NO 15, SEQ ID NO 16, SEQ ID NO 17, SEQ ID NO 18, SEQ ID NO 19, SEQ ID NO 20, SEQ ID NO 21, SEQ ID NO 22, SEQ ID NO 23, SEQ ID NO 24, SEQ ID NO 25, SEQ ID NO 26, SEQ ID NO 27, SEQ ID NO 28, SEQ ID NO 29, SEQ ID NO 30, SEQ ID NO 31, SEQ ID NO 32, SEQ ID NO 33, SEQ ID NO 34, SEQ ID NO 35, SEQ ID NO 36, SEQ ID NO 37, SEQ ID NO 38, SEQ ID NO 39, SEQ ID NO 40, SEQ ID NO 41, SEQ ID NO 42, SEQ ID NO 43, SEQ ID NO 44, SEQ ID NO 45, SEQ ID NO 46, SEQ ID NO 47, SEQ ID NO 48, SEQ ID NO 49, SEQ ID NO 50, SEQ ID NO 51, SEQ ID NO 52, SEQ ID NO 53, SEQ ID NO 54, SEQ ID NO 131, SEQ ID NO 132, SEQ ID NO 133, SEQ ID NO 134, SEQ ID NO 135, SEQ ID NO 136, SEQ ID NO 137, SEQ ID NO 138, SEQ ID NO 139, SEQ ID NO 144 which is a non-coding RNA.

In another embodiment the non-coding RNA is selected from the group consisting of pre-miRNA, pri-miRNA, miRNA, snRNA.

In another embodiment, the isolated nucleic acid comprises a sequence sharing at least 90 % identity with that set forth in the group consisting of SEQ ID NO 1, SEQ ID NO 2, SEQ ID NO 3, SEQ ID NO 4, SEQ ID NO 5, SEQ ID NO 6, SEQ ID NO 7, SEQ ID NO 8, SEQ ID NO 9, SEQ ID NO 10, SEQ ID NO 11, SEQ ID NO 12, SEQ ID NO 13, SEQ ID NO 14, SEQ ID NO 15, SEQ ID NO 16, SEQ ID NO 17, SEQ ID NO 18, SEQ ID NO 19, SEQ ID NO 20, SEQ ID NO 21, SEQ ID NO 22, SEQ ID NO 23, SEQ ID NO 24, SEQ ID NO 25, SEQ ID NO 26, SEQ ID NO 27, SEQ ID NO 28, SEQ ID NO 29, SEQ ID NO 30, SEQ ID NO 31, SEQ ID NO 32, SEQ ID NO 33, SEQ ID NO 34, SEQ ID NO 35, SEQ ID NO 36, SEQ ID NO 37, SEQ ID NO 38, SEQ ID NO 39, SEQ ID NO 40, SEQ ID NO 41, SEQ ID NO 42, SEQ ID NO 43, SEQ ID NO 44, SEQ ID NO 45, SEQ ID NO 46, SEQ ID NO 47, SEQ ID NO 48, SEQ ID NO 49, SEQ ID NO 50, SEQ ID NO 51, SEQ ID NO 52, SEQ ID NO 53, SEQ ID NO 54, SEQ ID NO 131, SEQ ID NO 132, SEQ ID NO 133, SEQ ID NO 134, SEQ ID NO 135, SEQ ID NO 136, SEQ ID NO 137, SEQ ID NO 138, SEQ ID NO 139, SEQ ID NO 144 such as 90 % identity, 91 % identity, 92 % identity, 93 % identity, 94 % identity, 95 % identity, 96 % identity, 97 % identity, 98 % identity, or 99 % identity.

Sequence identity

As commonly defined "identity" is here defined as sequence identity between genes or proteins at the nucleotide or amino acid level, respectively.

Thus, in the present context "sequence identity" is a measure of identity between proteins at the amino acid level and a measure of identity between nucleic acids at nucleotide level. The protein sequence identity may be determined by comparing the amino acid sequence in a given position in each sequence when the sequences are aligned. Similarly, the nucleic acid sequence identity may be determined by comparing the nucleotide sequence in a given position in each sequence when the sequences are aligned. To determine the percent identity of two nucleic acid sequences or of two amino acids, the sequences are aligned for optimal comparison purposes (e.g., gaps may be introduced in the sequence of a first amino acid or nucleic acid sequence for optimal alignment with a second amino or nucleic acid sequence). The amino acid residues or nucleotides at corresponding amino acid positions or nucleotide positions are then compared. When a position in the first sequence is occupied by the same amino acid residue or nucleotide as the corresponding position in the second sequence, then the molecules are identical at that position. The percent identity between the two sequences is a function of the number of identical positions shared by the sequences (i.e., % identity = # of identical positions/total # of positions (e.g., overlapping positions) x 100). In one embodiment the two sequences are the same length.

One may manually align the sequences and count the number of identical nucleic acids or amino acids. Alternatively, alignment of two sequences for the

determination of percent identity may be accomplished using a mathematical algorithm. Such an algorithm is incorporated into the NBLAST and XBLAST programs of (Altschul et al. 1990). BLAST nucleotide searches may be performed with the NBLAST program, score = 100, wordlength = 12, to obtain nucleotide sequences homologous to a nucleic acid molecules of the invention. BLAST protein searches may be performed with the XBLAST program, score = 50, wordlength = 3 to obtain amino acid sequences homologous to a protein molecule of the invention. To obtain gapped alignments for comparison purposes, Gapped BLAST may be utilised. Alternatively, PSI-Blast may be used to perform an iterated search which detects distant relationships between molecules. When utilising the NBLAST, XBLAST, and Gapped BLAST programs, the default parameters of the respective programs may be used. See http://www.ncbi.nlm.nih.gov. Alternatively, sequence identity may be calculated after the sequences have been aligned e.g. by the BLAST program in the EMBL database (www.ncbi.nlm.gov/cgi-bin/BLAST). Generally, the default settings with respect to e.g. "scoring matrix" and "gap penalty" may be used for alignment. In the context of the present invention, the BLASTN and PSI BLAST default settings may be advantageous.

The percent identity between two sequences may be determined using techniques similar to those described above, with or without allowing gaps. In calculating percent identity, only exact matches are counted. Sensitivity

As used herein the sensitivity refers to the measures of the proportion of actual positives which are correctly identified as such - in analogy with a diagnostic test, i.e. the percentage of sick people who are identified as having the condition.

Usually the sensitivity of a test can be described as the proportion of true positives of the total number with the target disorder. All patients with the target disorder are the sum of (detected) true positives (TP) and (undetected) false negatives (FN).

Specificity

As used herein the specificity refers to measures of the proportion of negatives which are correctly identified - i.e. the percentage of well people who are identified as not having the condition. The ideal diagnostic test is a test that has 100 % specificity, i.e. only detects diseased individuals and therefore no false positive results, and 100 % sensitivity, i.e. detects all diseased individuals and therefore no false negative results.

For any test, there is usually a trade-off between each measure. For example in a manufacturing setting in which one is testing for faults, one may be willing to risk discarding functioning components (low specificity), in order to increase the chance of identifying nearly all faulty components (high sensitivity). This trade-off can be represented graphically using a ROC curve.

Selecting a sensitivity and specificity it is possible to obtain the optimal outcome in a detection method. In determining the discriminating value distinguishing subjects or individuals having or developing e.g. colorectal cancer, the person skilled in the art has to predetermine the level of specificity. The ideal diagnostic test is a test that has 100% specificity, i.e. only detects diseased individuals and therefore no false positive results, and 100% sensitivity, i.e. detects all diseased individuals and therefore no false negative results. However, due to biological diversity no method can be expected to have 100% sensitive without including a substantial number of false negative results.

The chosen specificity determines the percentage of false positive cases that can be accepted in a given study/population and by a given institution. By decreasing specificity an increase in sensitivity is achieved. One example is a specificity of 95% which will result in a 5% rate of false positive cases. With a given prevalence of 1% of e.g. colorectal cancer in a screening population, a 95% specificity means that 5 individuals will undergo further physical examination in order to detect one (1) cancer case if the sensitivity of the test is 100%.

The cut-off level could be established using a number of methods, including :

percentiles, mean plus or minus standard deviation(s); multiples of median value; patient specific risk or other methods known to those who are skilled in the art.

Sample

In the present context, the term "sample" relates to any liquid or solid sample collected from an individual to be analyzed. Preferably, the sample is liquefied at the time of assaying.

In another embodiment of the present invention, a minimum of handling steps of the sample is necessary before measuring the expression of a RNA/cDNA. In the present context, the subject "handling steps" relates to any kind of pre-treatment of the liquid sample before or after it has been applied to the assay, kit or method. Pre-treatment procedures includes separation, filtration, dilution, distillation, concentration, inactivation of interfering compounds, centrifugation, heating, fixation, addition of reagents, or chemical treatment.

In accordance with the present invention, the sample to be analyzed is collected from any kind of mammal, including a human being, a pet animal, a zoo animal and a farm animal.

In yet another embodiment of the present invention, the sample is derived from any source such as body fluids.

Preferably, this source is selected from the group consisting of milk, semen, blood, serum, plasma, saliva, faeces, urine, sweat, ocular lens fluid, cerebral spinal fluid, cerebrospinal fluid, ascites fluid, mucous fluid, synovial fluid, peritoneal fluid, vaginal discharge, vaginal secretion, cervical discharge, cervical or vaginal swab material or pleural, amniotic fluid and other secreted fluids, substances, cultured cells, and tissue biopsies from organs such as the brain, heart and intestine. One embodiment of the present invention relates to a method according to the present invention, wherein said body sample or biological sample is selected from the group consisting of blood, faeces, urine, pleural fluid, oral washings, vaginal washings, cervical washings, cultured cells, tissue biopsies, and follicular fluid.

Another embodiment of the present invention relates to a method according to the present invention, wherein said biological sample is selected from the group consisting of blood, plasma and serum.

In a presently preferred embodiment of the present invention relates to a method according to the present invention, wherein said biological sample is serum.

The sample taken may be dried for transport and future analysis. Thus the method of the present invention includes the analysis of both liquid and dried samples.

Test sample

The test sample as used herein refers to a RNA/cDNA sample, and can be of any source.

Reference

As used herein can a reference refer to a reference sample or a reference subject. Reference sample

The reference sample can consist of one or more RNA/cDNA samples, and can be of any source.

In some embodiments is the reference another gene or an intragenetic reference such as an exon within the gene and/or RNA transcript variant of interest.

In an embodiment of the present invention is the expression of one or more specific exons in the RNA transcript variants used as reference.

In a more specific embodiment of the present invention are these specific exons in the RNA transcript variants exon 1, exon 2, exon 3, exon5, exon 6, exon 7, exon 8 or exon 9 for SLC39A14 and exon 1, exon 2, exon 3, exon 4', exon5, exon 6, exon 7 for VNN1.

The genetic boundaries of the exons can be found in the examples and tables of the present application.

In some embodiments the reference sample is from the same species as the comparable test sample. The reference sample can be obtained as an average expression from 1 to ~ n number of samples. The reference sample can also reflect a pool of reference samples.

Test subject

As used herein refers a test subject to the subject from which the test sample is obtained.

In accordance with one embodiment of the present invention, the sample to be analyzed may be collected from any kind of mammal, including a human being, a pet animal, a zoo animal and a farm animal.

Reference subject

As used herein a reference subject refers to the mammal from which the reference sample is obtained.

The reference subject can be obtained as an average from 1 to ~ n number of subjects or seen as a population.

General

It should be noted that embodiments and features described in the context of one of the aspects of the present invention also apply to the other aspects of the invention. All patent and non-patent references cited in the present application, are hereby incorporated by reference in their entirety.

The invention will now be described in further details in the following non-limiting examples.

Examples

Materials and methods

Colorectal cell lines and tissue samples.

The project involved analyses of six colon carcinoma cell lines (HT29, HCT15, SW48, SW480, RKO, and LS1034) from which RNA was isolated by Trizol

(Invitrogen, Carlsbad, California, USA). Ten primary colorectal carcinoma samples and ten normal colorectal samples from cancer patients were also included, from which RNA was isolated by the All prep DNA/RNA mini kit (Qiagen) and the

Ribopure™ kit (Applied Biosystems/Ambion, Foster City, California, USA).

Publicly available databases

Sequence information about genes and their different transcripts have been investigated using the EnsembI genome browser and all herein described sequences are in compliance with release 50, published July 2008. Sequence specificities, on the other hand, have been assessed by BLAST. These searches were carried out in the human genomic plus transcript database, by use of the nucleotide blast program, and the megablast algorithm.

Exon microarray analysis

The GeneChip® Human Exon 1.0 ST Array (Affymetrix, Santa Clara, CA, USA) provides genome-wide detection of RNA expression at both gene and exon levels. The microarray has approximately 5.4 million probes grouped into 1.4 million probesets examining more than a million known and predicted exons. The probes are distributed in the different exons along the entire transcript length, and for a gene with ten exons, there are roughly 40 probes matching its sequence. With probes in different exons along the transcript it is possible to monitor the level of expression for each exon compared with the others in the gene and thereby detect different transcript variants created after events such as alternative splicing and alternative promoter usage or poly-adenylation sites.

Ten normal colonic tissue samples, ten colorectal cancer tissue samples and six colorectal cancer cell lines (HT29, HCT15, SW48, SW480, RKO, and LS1034) were analysed. Raw data were imported into the XRAY software (version 2.81; Biotique Systems Inc., Reno, Nevada, USA) where quantile normalisation and calculation of probeset expression values were performed and summarized. Only "core" probesets (RefSeq and full-length GenBank mRNAs) were analysed and the expression score for a probeset was defined to be the median of its probe expression scores. For each probeset the log2-ratio of expression level in test samples to that observed in control samples were calculated.

Exon microarray data were investigated from genes resulting from all the three different input strategies (outlier expression profiles, known and putative fusion genes, and ETS family members). The longitudinal exon expression profile along the entire transcript length of each gene was visualized by an in-house created visual basics script, and evaluated manually by looking for profiles where individual samples were overexpressed only in the 3' part of the transcript compared to the rest of the samples (examples in Figure 4 and Figure 8). Genes with this type of profile were investigated further in the laboratory with 5'-RACE, cloning and sequencing.

Rapid Amplification of cDNA Ends

The complete 5'- and 3'-ends of cDNA can be amplified by PCR, using a technique variously called rapid amplification of cDNA ends (RACE), one-sided PCR and anchored PCR. The technique uses PCR to amplify partial cDNAs that represent the region between the 5'- or 3'-end and a single point in an mRNA transcript. The main requirement is that a short stretch of sequence in the mRNA of interest is known. A gene-specific primer (GSP), oriented in the direction of either the 5'- or 3'-end, is designed to anneal in the already known sequence. Extension of the cDNA from the end and back to the known region is achieved by using a primer annealing to the pre-existing poly(A) region (3'-RACE) or to an appended homopolymer tail or linker (5'-RACE). 5'-RACE

In this project 5'-RACE was performed using the SMART RACE cDNA Amplification kit (Clontech, Mountain View, California, USA). The first-strand synthesis is primed with an oligo-(dT) primer and performed by a Moloney murine leukemia virus reverse transcriptase (MMLV RT) which adds 3-5 residues (predominantly cytosines) upon reaching the 3'-end of the first-strand cDNA. A SMART II A oligo in the reaction mix contains a terminal stretch of G-residues which anneals to this cDNA tail. MMLV RT switches template from the mRNA to the SMART oligo and generates a complete cDNA copy of the mRNA with the additional SMART sequence at the end. MMLV RT's terminal transferase activity is most efficient when the enzyme has reached the end of the RNA-template and the SMART sequence is therefore typically added only to complete first-strand cDNAs.

The 5'-end of the cDNA can then be amplified using a universal primer (UP) which anneals in the SMART sequence and a primer specific for the gene of interest. The GSP must be between 23 and 25 nucleotides long, have a GC-content between 50 and 70 percent, and an annealing temperature above 70°C.

On occasion, a reverse transcription reaction can be non-specifically primed and result in a cDNA containing the SMART sequence at both ends. To reduce the likelihood of such aberrant products, a mixture of long and short UPs (with excess of the short UP) is used. The long UP contains inverted repeat elements. During PCR of a cDNA with SMART sequence in both ends, the long UP will anneal in both ends and the inverted repeats anneal to each other, making a panhandle-like structure. This blocks amplification of such aberrant products because the short UPs are unable to anneal.

Generation of 5'-RACE-ready cDNA was performed using the SMART RACE cDNA amplification kit (Clontech) and PrimeScript reverse transcriptase (Takara Bio Inc., Otsu, Shiga, Japan). One pg total RNA was combined with 2.4 μΜ oligo-(dT) primer, 2.4 μΜ SMART II A oligo, and sterile water to a total volume of 5 μΙ. The reaction mix was first incubated at 70°C for 2 min to allow the primers to anneal and then on ice for two minutes before adding 1 x first-strand buffer, 2 mM dithiothreitol (DTT), 1 mM dNTP, and 200 U PrimeScript reverse transcriptase to a total volume of 10 μΙ. Elongation of the cDNA at 42°C for 90 min followed. The first-strand reaction was then diluted in 100 μΙ Tricine-EDTA buffer and the reaction was stopped by incubation at 72°C for 7 min.

RACE reactions were performed using the SMART RACE cDNA amplification kit and the Advantage 2 PCR kit (Clontech). 1 x Advantage 2 PCR buffer, 0.2 mM dNTP mix, IX Advantage 2 PCR polymerase mix, 2.5 μΙ RACE-ready cDNA, 1 x Universal primer mix (UPM), 0.2 μΜ GSP, and PCR-grade water was combined to a final volume of 50 μΙ. The cycling conditions were as described in Table 1.

Nested RACE was then performed by combining the same reagents as for RACE, but this time with 5 μΙ diluted RACE product as template and nested primers. The nested RACE was run by 25 cycles of 30 sec at 94°C, 30 sec at 68°C, and 3 min at 72°C.

Cloning

Cloning and transformation was performed using the TOPO TA Cloning Kit

(Invitrogen). This kit takes advantage of topoisomerase I and the fact that it can bind to DNA and cleave the phosphodiester backbone after 5'-CCCTT-3'. The energy from the broken bond is conserved by formation of a covalent bond between the cleaved strand and the topoisomerase I. Before cloning, the vector is cut into linear form, with single 3' thymidine (T) overhangs. Taq polymerase has a non-template dependent terminal transferase activity, which adds a single deoxyadenosine (A) to the 3'-ends of PCR products. By reversing the cleavage reaction the PCR product with its A-overhang is readily incorporated into the T-overhang containing vector and the topoisomerase is released.

The vector contains the lethal ccdB gene fused to the LacZa gene. Ligation of the PCR product disrupts expression of the ccdB-LacZa gene and allows only positive recombinants to grow. A gene for ampicillin resistance in the vector ensures that only transformed bacteria will grow in the presence of this antibiotic compound.

Four μΙ PCR product eluted from an agarose gel was mixed with 1 μΙ salt solution and 1 μΙ TOPO vector before incubation at room temperature for 30 min. The cloning reaction was then transferred to ice. Two μΙ of the reaction was transferred to a vial of One Shot TOP10 E. coli and incubated on ice for 5-30 min. The cells were given a heat shock for 30 sec at 42°C and immediately transferred back to ice. 250 μΙ of room temperature S.O.C. medium was added and the cells incubated horizontally at 37°C and 200 rpm for 1 h. After the incubation 50 μΙ and 75 μΙ of the transformation mix was spread on pre-warmed selective LB plates containing 100 pg/ml ampicillin. The plates were incubated over night at 37°C.

Individual colonies were picked from selective plates and used to inoculate individual cultures consisting of 5 ml LB-medium and 10 μΙ ampicillin. The cultures were incubated at 37°C and 250 rpm over night. Bacterial cells were then harvested by centrifugation and plasmid DNA was purified using the QIAprep Spin Miniprep kit (Qiagen).

DNA Sequencing

The sequencing reaction was performed in a 96-well Optical Reaction Plate and consisted of purified template DNA (either PCR product eluted from agarose gel or plasmid DNA from Miniprep purification), primer (forward or reverse), BigDye Terminator v3.1 or vl. l premix (Applied Biosystems), BigDye Sequencing buffer (Applied Biosystems) and Milli-Q water to a total volume of 10 μΙ. First, the reaction mixes were incubated at 96°C for 2 min, followed by 25 thermal cycles of 15 sec at 96°C, 5 sec at 50°C, and 4 min at 60°C. The thermal cycling was performed on an MJ Research Cycler (BIO-RAD).

The BigDye Terminator v3.1 premix was used when the fragment to be sequenced were longer than 500 base pairs and the vl. l for shorter fragments. The premix contains dNTPs and ddNTPs. The different ddNTPs are modified with fluorescent labels which emit light at specific wavelengths when exposed to a laser beam. This makes it possible to visualise the different bases.

Product purification

After the sequencing reaction unincorporated dye terminators, salts and other charged molecules must be removed. This was done by using the BigDye

Xterminator Purification Kit (Applied Biosystems). Forty-five μΙ of SAMTM solution and 10 μΙ of XterminatorTM were added to the sequencing reaction after completion of thermal cycling. The reaction mixes were then vortexed for 30 min and briefly centrifuged in the end. The SAM solution enhances the performance of the Xterminator solution and stabilises the post-purification reactions. The Xterminator, on the other hand, scavenges unincorporated dye terminators and free salts.

Capillary analysis

The 96-well Optical Reaction Plate was sealed with a 3100 Genetic Analyzer Plate Septa (Applied Biosystems), placed in a 96-well Plate Base, and inserted into a fully automated AB 3730 DNA analyser (Applied Biosystems). Inside the analyser the 48- capillary array is filled with POP7 polymer (Applied Biosystems). The samples are then loaded and separated according to size as they migrate through the polymer- filled capillaries. As the fluorescently labelled DNA fragments reach the detection window, a laser beam excites the dye molecules and causes them to fluoresce. The Data Collection software reads and interprets the fluorescence data before displaying them as an electropherogram. The samples were analysed using the software Sequencing Analysis 5.2 (Applied Biosystems), and all electropherograms were read both manually and automatically.

Real time RT-PCT

The cDNA synthesis was performed using the same kit as previously described. The pre-designed commercial quantitative RT-PCR assay was carried out in a fast optical 96-well reaction plate (Applied Biosystems), and the custom-designed assays were performed in standard 96- or 384-well optical reaction plates (Applied Biosystems). Different TaqMan master mixes, reaction volumes, and thermal cycling conditions were used with regard to whether the reactions should be carried out in fast or standard, or 384- or 96-well plates. The TaqMan Fast Universal PCR Master Mix (No AmpErase UNG, Applied Biosystems) was utilized in fast reactions, and the TaqMan Universal PCR Master Mix (AmpErase UNG, Applied Biosystems) was used in standard reactions. The final concentrations of master mix, forward and reverse primers, and probe in the standard reactions were 1 x, 0.9 μΜ of each, and 0.2 μΜ, respectively. In the fast reaction, the end concentrations of master mix and TaqMan Gene Expression Assay (primers and probe) were both 1 x. A total reaction volume of 20 μΙ was used when the reactions were performed in 384- and fast 96-well plates, as distinct from standard 96-well plates, where the total volume per reaction was set to 25 μΙ. RNase free water (Sigma-Aldrich) was added to the correct total volume. In each setup, the amount of starting material (cDNA) was 10 ng. Multiplex real-time RT-PCR was done as well, adding one primer set, and two different probes with distinct dyes (FAM and VIC), in the same well. In these cases, the concentration of each of the primers where doubled compared to a standard setup, whereas the remainder (plate, master mix, final concentrations, volume, and thermal cycling conditions) was the same as for a standard assay.

The plates were incubated, and fluorescence measured, on an ABI 7900HT Fast Real-Time PCR System (also known as a "TaqMan"; Applied Biosystems). The thermal cycling conditions differed in the fast and standard reactions (see below).

Thermal cycling conditions real-time RT-PCR.

Standard Fast

UNG activation : 50 °C 2 min

Polymerase activation : 95 °C 10 min 20 sec

Denaturation : 95 °C 15 sec 1 sec

Annealing and extension : 60 °C 1 min 20 sec

Number of cycles: 40 40

The pipetting robot EpMotion 5075 (Eppendorf, Hamburg, Germany) was used to pipette template to the wells in 384 plates, but the 96-well plates were set up manually. Master mix was distributed manually with a multi-channel pipette.

A standard curve was produced by serially diluted universal human reference (UHR) cDNA, synthesised from UHR RNA (Stratagene), of known concentrations. All samples were run in triplicates, and the endogenous control gene assay ACTB (Applied Biosystems) was performed on all the samples.

Example 1 Identification of novel transcripts

Three starting points were used for the candidate gene selection in the hunt for fusion genes; genes with outlier expression profiles, known and putative 3' fusion gene partners and members of the ETS gene family.

Here, 508 genes (131 outliers, 349 known and putative fusion genes and 28 ETS family members) were investigated with the exon microarray. Eleven genes

(RAD51L1, NKAIN2, VNNl, C4BPB, HOXCll, TFR2, SERPINB7, TFPT, GJB6, PRRXl, and PRRX2) had a longitudinal profile along the exons where one or two of the cell lines deviated from the rest only in the 3'-end. Five of these genes (TFR2, SERPINB7, C4BPB, VNN1, and GJB6) had outlier expression profiles in colorectal tissues, and the other six genes (PRRX1, PRRX2, NKAIN2, HOXC11, TFPT, and RAD51L1) are known fusion gene partners. None of the ETS family members and none of the putative fusion genes exhibited the desirable profile. For each of the 11 genes 5'-RACE and nested RACE was performed (see Figure 1 for representative results). Products were separated with gel electrophoresis, cut and eluted from the gel, cloned, and sequenced. No fusion genes were found from analysis of these genes, but novel transcript variants were found in all of the 11 genes.

The exon expression profile of RAD51L1 in the SW48 cell line deviated from the other cell lines by having higher expression from exon seven and throughout the gene (Figure 2A). Five transcript variants with a total of 14 exons are known for RAD51L1, but sequencing of the 5'-RACE products from SW48 revealed six novel transcript variants which all included novel exons located inside intron number seven (Figure 2B). The novel exons are spliced together in different ways to create the different transcripts. See Appendix II for details about each transcript and the different exons. The nucleotide sequences of the novel transcripts were evaluated by use of the Translate tool for translation of nucleotide sequences into protein sequences. This revealed that the transcripts B and F contain open reading frames (i. e., a start codon which is not followed by an immediate in-frame stop codon) of 66 amino acids, and these are thus potentially protein-coding.

The same type of exon expression profile was found for NKAIN2 in both a cell line (LS1034) and a primary tumour (C1033III). These profiles show a higher expression of exons eight, nine, and ten in LS1034 (Figure 3A) and C1033III (Figure 3B) compared to the other cell lines and primary tumours.

Three transcripts are known for NKAIN2, all of which are transcribed from the same promoter (Figure 3C). Sequencing of the 5'-RACE products from both LS1034 and C1033III reveals the presence of eight novel transcripts including four novel exons, here denoted α, β, γ, and δ. Exon a is used as first exon in transcripts A, D, E, and G whereas exon γ is the first exon in transcript B. Exons β and δ, on the other hand, are located downstream of exon eight and nine, respectively. In the different transcripts, transcription is initiated at exon a, four, y, nine, or ten. The Translate tool reveals transcripts A, G, D, F, and E as potentially protein-coding, with open reading frames of up to 173 amino acids, whereas transcripts C, B, and H probably are not.

The exon expression profile for VNN1 in the cell line HT29 deviated from that of the other cell lines by higher expression of exons six and seven (Figure 4A). Sequencing of the 5'-RACE products from VNN1 revealed three transcript variants in HT29 (Figure 4B). One transcript variant with seven exons is known for VNN1, but exons one to five in this transcript were never detected in HT29, instead two new exons, a and β, located inside intron number five are present. Transcript A consists of exon a followed by exon β and exon six. The Translate tool indicates that the transcript might encode a protein of 83 amino acids. Transcript B is quite similar to A, but with a 35 basepairs longer exon β. This results in frame shift from the subsequent exon of transcript A, introducing a stop codon, and B is therefore most likely non- coding. In transcript C a short exon a is directly followed by exon six. The Translate tool revealed no open reading frame from this sequence.

The exon expression profile for C4BPB in C1034III deviated from the other primary tumours by higher expression from the middle of the second exon and throughout the gene (Figure 5A). Five different transcripts, transcribed from two different promoters, are known for C4BPB (Figure 5B). Three different transcripts were found by sequencing of the 5'-RACE products from C1034III, all of which seem to be transcribed from the two known promoters (Figure 5B). Transcript A consists of the reference exon one and an enlarged exon two with additional sequences 5' to the reference exon. Transcript B starts in exon two, in accordance with both

ENST00000243611 and ENST00000367076. Transcript C is similar to

ENST00000367078, but with a larger first exon. Since the gene-specific primer is located relatively close to the 5'-end of the gene, we do not have enough

information on whether the two new transcripts, A and C, are protein-coding.

The exon expression profile for HOXC11 in the primary tumour C1402III deviates from the profile of the other tumours with higher expression from the end of exon one and throughout the gene (Figure 6A). One transcript with two exons is known for HOXC11 (Figure 6B). Sequencing of the 5'-RACE products revealed two novel transcripts in C1402III (Figure 6B). These transcripts consist of a novel exon, here denoted a, of variable length, spliced to exon two in the known transcript. The Translate tool indicates that transcript A, with the large exon a, exhibits an open reading frame encoding up to 119 amino acids with multiple possible initiation codons. The C-terminal end of the putative peptide generated from transcript A is identical to the C-terminal end of the peptide generated from ENST00000243082. Transcript B has a short exon a and only a quite short open reading frame encoding 38 amino acids, identical to the last part of the open reading frame in transcript A.

Two cell lines, RKO and SW48, had similar exon expression profiles for TFR2. These profiles deviated from those seen in the other cell lines by higher expression of exon eight and throughout the gene (Figure 7A). One transcript with 18 exons is known for TFR2, and sequencing of the 5'-RACE products from RKO and SW48 revealed ten novel transcripts (Figure 7B). Exons one, two, and three were never present in these transcripts, and instead, all transcripts were initiated from exons four, six, and seven. The transcripts differ with regard to the amount of intron sequence included around the known exons. The Translate tool indicates an open reading frame in transcripts A, E, F, and H encoding 46 amino acids and an open reading frame encoding 160 amino acids in transcript D. For all these five

transcripts, no stop codon is encoded and the open reading frame continues into the exon(s) downstream of the primer location. No open reading frames were found for transcripts B, C, G, I, and J.

The exon expression profile of SERPINB7 in the LS1034 cell line deviated from the other cell lines in exons five to nine (Figure 8A). Two transcript variants are known for SERPINB7 with a total of nine exons, where the first two are non-coding (Figure 8B). Sequencing of the 5'-RACE products revealed three variants in LS1034.

Transcript B exhibits a novel first exon located inside intron number two. The Translate tool indicates that the transcript variant encodes the same protein as the two known transcripts, but has a different 5'-UTR. Transcript A is identical to ENST00000398019. Transcript C only includes exons four to six and the Translate tool reveals that no open reading frame is encoded by the transcript.

The exon expression profile for TFPT in SW48 shows higher expression in exons four, five, six, and seven compared to the other cell lines (Figure 9A). Four transcripts, transcribed from three different promoters and with a total of seven exons, are known for TFPT (Figure 9B). Sequencing of the 5'-RACE products revealed the presence of two transcripts in SW48 (Figure 9B). Transcript A is transcribed from exon three and the Translate tool indicates that no open reading frame is encoded by the transcript. Transcript B, on the other hand, is similar to one of the known transcripts (ENST00000301757), but with a larger first exon.

The exon expression profile for GJB6 in HT29 deviated from the other cell lines by having higher expression in exons five and six (Figure 10A).

Four transcripts with a total of six exons are known for GJB6. Sequencing of the 5'- RACE products revealed the presence of six transcript variants in HT29 (Figure 10B). Transcript A only includes the last exon, and do not encode an open reading frame. Transcripts B and C, are identical to two of the known protein-coding variants (ENST00000400066 and ENST00000400065, respectively). Transcript D presents the same exon composition as ENST00000400066 but the sequence of exon five is 21 basepairs longer on its 5'-end, which induces seven new amino acids upstream of the coding region. Transcript E and F are initiated in exons two and five, respectively, and the Translate tool indicates that they encode an intact protein, but have a different 5'-UTR.

The exon expression profile for PRRX1 revealed higher expression of exons two to five in SW48 as compared to the other cell lines (Figure 11A). Two transcripts with a total of five exons are known for PRRX1, and sequencing of the 5'-RACE products from SW48 revealed nine transcript variants with a total of five novel exons localised in the 3'-end of intron one (Figure 11B). Exon one is not present in any of the transcripts, and instead, transcription is initiated at exons α, y, and δ. The novel exons are spliced together in multiple ways to create the nine different transcript structures identified. The Translate tool indicates the presence of open reading frames in transcripts A and B which might encode up to 83 amino acids. No stop codons were found in these frames, indicating the presence of more coding exon(s) 3' of the primer location. None of the other transcripts seem to contain open reading frames.

The exon expression profile for PRRX2 in the primary tumour sample C1033III deviates from the other samples by having higher expression in the last exon of the gene (Figure 12A). One transcript, consisting of four exons, is known for PRRX2, and sequencing of the 5'-RACE products revealed two novel transcript variants, A and B (Figure 12B). Transcript A includes parts of exon three spliced to exon four, whereas transcript B only consists of exon four. Eleven clones exhibited transcript A, and transcription was initiated at the exact same location for all clones (Appendix II). The Translate tool indicates that none of the transcripts are protein-coding.

Methodological considerations

Three starting points were used for the candidate gene selection in the search for novel transcript variants, including a search for fusion genes; genes with outlier expression profiles, known and putative 3' fusion gene partners and members of the ETS gene family. A fusion gene usually leads to the overexpression of the

downstream fusion partner and a fusion gene is usually only present in a subset of cancer samples. The formation of a fusion gene therefore leads to overexpression of the downstream partner gene in only some of the samples, giving rise to an outlier expression profile. Previously, cancer outlier profile analysis has been used to calculate outlier profiles in the search for novel fusion genes (Tomlins et al., Science 2005). Known and putative 3' fusion gene partners and ETS gene family members were included because of their known susceptibility for undergoing rearrangements and because the same fusion genes (and in particular the same fusion gene partners) can be present in different cancer types.

Analysis of the longitudinal exon expression profile turned to be an important step in the process of enriching for genes with alterations in their transcript structures. If every exon in a gene is under the control of the same promoter, we would expect the exon expression levels to be similar throughout the gene. If, on the other hand, a gene has a second promoter or is the downstream partner in a fusion gene (and thus has downstream exons under the control of a new promoter), the exons downstream of the new promoter/breakpoint will be under the control of a different promoter than the upstream exons. The 5'-portion of the original gene is therefore regulated by one promoter and the 3'-portion by another, leading to different expression of the two parts. This may give rise to longitudinal exon expression profiles looking like the ones seen in Figure 2A to Figure 12A, where exons in the 3'-end of a gene have higher expression than the 5'-exons in certain samples as compared to others.

To investigate the transcript structure upstream of the altered exon expression, 5'- RACE was used. One debate concerning RACE methods is whether the entire beginning of the transcript is reached. For the SMART RACE kit used in the present project, it has been reported that 70-90 % of the products correspond to the actual 5'-end of the mRNA. The majority of transcripts found in this project may therefore be considered to include the 5'-end of the mRNA. This is also supported by findings shown in Appendix II, where different clones for the same transcript start at the exact same base, indicating that this is the first base to be transcribed into mRNA. An example can be seen in Appendix II for PRRX2 transcript A, where all eleven clones started with the same nucleotide.

Multiple transcripts are found for the majority of genes. The gene-specific primers used in the RACE setup anneal to a particular exon. By use of the exon microarray expression profiles, gene-specific primers could be designed to anneal in exons indicated to be highly expressed, and therefore most likely also included in a potential novel transcript variant initiated from a novel and strong promoter.

Two steps in the process from mRNA to sequenced 5'-cDNA ends were essential for success: Firstly, since the RACE method only applies one gene-specific primer, it is necessary to perform nested RACE with a nested gene-specific primer to ensure gene-specific RACE products. Secondly, it is necessary to separate nested RACE products on an agarose gel, followed by elution of individual bands, prior to cloning. Abrogating this step will favour cloning of short products. Some of the adenosine overhangs produced by the PCR reaction, and necessary for cloning into the TOPO vectors, are lost during the gel elution step, thus making the cloning reaction less effective. Accordingly, the amount of transformation mix had to be increased to ensure sufficient growth of transformed bacteria.

Novel exons and transcripts

Among the 11 genes investigated in the laboratory because of exon expression profiles deviating in the 3'-end of the transcripts, five were initially included as candidate genes due to outlier expression profiles in tissues from colorectal cancer (TFR2, SERPINB7, C4BPB, VNN1, GJB6) and six due to their known participation as fusion gene partners in other cancer types (RAD51L1, NKAIN2, HOXC11, TFPT, PRRX1, and PRRX2). In total, laboratory investigations of the 11 genes lead to the discovery of 57 novel transcript variants, including 22 novel exons and 34 putative novel promoters in colorectal cancer. In the following each gene and its transcript variants will be discussed in more detail. Large discrepancies are seen in different human genome databases with regards to, for instance, what is considered a transcript variant and the nomenclature of exons and transcripts. Therefore, throughout the project one genomic database, Ensembl, have been used to asses the different transcripts and exons known for a given gene. Ensembl, which is curated by the European Bioinformatics Institute, is considered a comprehensive, well-annotated and stable database, where annotated genes and transcripts are based on mRNA and protein sequences deposited into public databases from the scientific community.

For RAD51L1, the transcription start sites of the herein identified novel transcript variants indicate the presence of three novel promoters, at exons denoted α, β, and y. The exon expression profile for RAD51L1 (Figure 2) shows higher expression of the last exons in the investigated cell line as compared to the others and therefore indicate that one or both of the alternative promoters are more activated than the reference promoters. The investigated cell line, SW48, also has higher expression of exon two compared to the other cell lines. This can not be explained by the transcripts described in this project because exons one to seven are not present in any of them. The high expression in exon number two might be explained by transcripts which do not contain exon eight, and therefore are not detected with the RACE primed for this exon.

For NKAIN2, the novel exon a is used as first exon in four of the sequenced transcripts and indicate the presence of a novel promoter. Promoters might also be present at exons four, y, nine and ten, as these are the first exons in the other four transcripts. The exon expression profiles of the cell line and tumour sample investigated deviate most strikingly from the other cell lines and tumour samples in exon eight, nine, and ten. In addition, they both also have the highest expression in exon five, as compared to samples of the same kind, which is in line with the presence of this exon in five transcripts.

The exon expression profile for VNN1 in HT29 was quite striking (Figure 4) with the higher expression of exons six and seven as compared to the average expression of the ten tumour samples, which are somewhat upregulated compared to normal samples and cell lines. Three transcript variants with two novel exons were found. For all novel transcript variants of VNN1 expression starts in the novel exon a, indicating the presence of a novel promoter. To account for the high expression of exons six and seven, the promoter used to generate these transcripts must be more active than the normal promoter in VNN1.

The enlarged exon two seen in transcript A of C4BPB might constitute a longer 5'- UTR and thereby affect its stability and/or regulation of translation. Transcript C might be the same as ENST00000367078. The first exon is bigger in transcript C, but this might be due to use of different TSSs and thus, the promoter is not necessarily a novel one.

Both of the novel transcripts seen for HOXCll consist of a version of exon a, spliced to exon two in the reference transcript. This indicates the presence of a novel promoter at exon a. The possible protein encoded by transcript A, might be a truncated version of the known protein product of ENST00000243082 or a novel protein with identical C-terminal end.

The novel transcript D seen in TFR2 consists of exons four to eight and was only found in the RKO cell line. The exon expression profiles for the two investigated cell lines deviate most from the other cell lines in exons eight to ten, but the presence of exon four in transcript D is in concordance with the peak seen at this position in the exon expression profile for RKO. The drop in expression seen for exon five for all cell lines might be due to a non-functioning probeset. All transcripts are initiated from either exon four, six, or seven, indicating the presence of novel promoters in these regions.

Two novel and one known transcripts were found for SERPINB7 (Figure 8). For SERPINB7, the first exon seen in transcript B is likely non-coding and can give the potentially encoded protein a different 5'-UTR than the known isoforms of the gene. This might affect the stability and regulation of the encoded protein.

The exon expression profile for TFPT in SW48 shows high expression of exon one, but lower expression of exons two and three. Exon two is not present in the two transcripts seen in SW48 and might therefore explain the drop in the expression profile. Exon three, on the other hand, is present in both transcripts. This drop in expression is seen, in various degrees, in this location for all the cell lines and may be due to a probeset not working properly. The enlarged first exon in transcript B might be due to alternative TSS use as compared to the known transcript, and not indicate the presence of a novel promoter. The entire coding region of GJB6 is located in exon 6. The enlarged fifth exon seen in transcript D alters the 5'-UTR and might therefore affect the stability and/or regulation of translation. Transcripts E and F differ from the reference transcripts and indicate the presence of new promoters in front of exons two and five, respectively. The potential proteins encoded by these transcripts are identical, but the transcripts exhibit different 5'-UTR as compared to the known proteins and might therefore be regulated differently. None of the transcripts sequenced from the HT29 cell line includes exon 3, thus explaining the drop seen at this position in the exon expression profile.

In the novel transcript variants seen for PRRX1, transcription is initiated at exons a, Y, and δ indicating the presence of three novel promoters. The exon expression profile for the investigated cell line shows continuous high expression of PRRX1 in exons three, four, and five. This indicates the presence of all these exons in the full- length transcripts and is in concordance with the lack of stop codons upstream of the primer location in transcripts A and B. To account for the elevated expression of exons two to five, one or more of the novel promoters found in the investigated cell line must be more active than the normal promoter for PRRX1.

Eleven clones containing transcript A of PRRX2 were sequenced, all of which were of the exact same length because transcription was initiated at the exact same nucleotide. This indicates that the far 5'-end of the transcripts were reached using 5'-RACE and therefore also supports the findings of a wider repertoire of promoters for the other genes investigated in this project.

The Translate tool used to translate nucleotide sequences to peptide sequences of potential proteins has been used to evaluate whether or not different transcripts have the possibility to be protein-coding. The transcripts referred to as non-coding have been of two types; either with many stop codons dispersed throughout the nucleotide sequence, in all three reading frames, or a transcript sequence with no start codon. The latter type was found in transcripts from TFR2, SERPINB7, TFPT, and GJB6. The nucleotide sequences from these transcripts were typically

containing an open reading frame, but did not include start codon for this frame. Nevertheless, these transcripts may as well represent sequences where the 5'-end of the cDNA has not been reached. True non-coding transcripts may as well be functionally relevant to the cells. Over the past few years, several long non-coding RNAs have been discovered. Many of these RNAs control the activity of protein-coding genes and do so in a variety of ways without necessarily being dependent on the exact sequence of the RNA. For example, as seen from the DHFR gene, a non-coding RNA generated from one promoter in a gene can regulate the transcription of protein-coding transcripts generated from another promoter within the same gene.

Nonsense-mediated mRNA decay represents a posttranscriptional process which selectively recognises and degrades mRNAs with truncated open reading frames. The novel transcripts detected in this project are clearly not degraded, as their corresponding genes were included in the study based on high mRNA levels. This is yet another indication that they may have functional implications to the cells.

The transcripts described in this example display 34 potentially novel promoters. This includes both transcripts potentially encoding the reference proteins but containing different 5'-UTR (as seen for GJB6, transcripts E and F) and transcripts potentially encoding novel proteins (as seen for RAD51L1, transcripts B and F). Heterogeneous 5'-UTRs can affect the stability and translation efficiency of the mRNAs and thereby affect the amount of protein present in a cell, whereas isoforms of the same gene may have different functions. The potential proteins encoded by transcripts identified in this project may therefore introduce effects to a cancer cell which are different to those of the proteins encoded by the reference transcripts.

As seen from Appendix II, the exact TSSs for the same type of transcripts within different clones differ by some nucleotides. This is in accordance with the findings that most human promoters lack one distinct TSS, but instead consist of a series of closely located TSSs spread over around 50 to 100 basepairs. For some transcripts, the TSSs seen in Appendix II are separated by more than 100 basepairs, and may therefore indicate the presence of more than one core promoter.

Summarised, the exon expression levels for 508 genes were investigated. Eleven of the genes had deviating exon expression profiles indicating qualitative changes in the transcript structure and were therefore investigated in the laboratory. No new fusion gene was found, but 57 novel transcript variants including 22 novel exons and 34 putative promoters were identified from colorectal cancer cell lines and tissue samples. Thus, in conclusion, we consider our novel strategy for identification of novel transcript variants in colorectal cancer as successful. The novel transcripts will be further investigated in our laboratory to elucidate their prevalence and clinical relevance in colorectal cancer, as well as their cancer-specificity.

Example 2

All three transcripts (VNNl A, B and C) originate partly from within the genomic portion annotated as intron 5, between exons 5 (ENSE00000764053) and 6

(ENSE00000764052), of the VNNl gene (ENSG00000112299; ENST00000367928). VNNl-intron 5 is located 133,005,645 to 133,013,361 basepairs from the p- telomere of chromosome 6 (Ensembl release 56). The VNNl gene is transcribed from the minus-strand; hence, the sequence starts further away from the p- telomere than it ends. The start and end positions of the transcripts can be found in Table-A-II-3.

Example 3

SLC39A14, also known as Zrt- and Irt-like protein 14 (ZIP14), is transcribed from the plus strand of cytogenetic band 8p21.3. The Ensembl Genome Browser, release 56, has annotated four transcript variants of SLC39A14.

The exon microarray data are shown in Figure 16A. From the exon-wise plot it seems that when exon four primed is expressed, the exon four is not expressed, and vice versa, which implies that the two exons four are mutually exclusive. This assumption is reinforced by the fact that in known transcripts from this gene, the exons four-primed and four never exist in the same transcript. As the two exons four have identical lengths, it is impossible to differentiate between the two splicing events by standard RT-PCR with primers in flanking constitutive exons.

Consequently, two real-time RT-PCR assays with exon specific probes were designed to validate the structure and quantities of the transcripts resulting from these mutually exclusive splicing events (Figure 16B).

In the first-line validation, these assays were performed on 19 of the same samples as examined by the exon microarray analysis, three normals (which were selected based on the exon-wise plots) and 16 CRC tissue samples, in addition to six colon cancer cell lines (Figure 17). This method revealed a clear difference between normal tissues and CRC samples (both tissues and cell lines). The real-time assay with a probe in exon four primed detected accumulation of PCR product in all three normal samples (all samples are run in triplicates), but only in two of 16 tumour samples and in none of the six cell lines. Furthermore, the real-time assay with a probe in exon four detected PCR product in all cell lines, in addition to 15 of 16 tumour samples, but in none of the normal tissue samples.

Subsequently, these transcripts were further investigated by expanding the sample series of clinical CRC and normal tissue samples. The Ct values obtained for each of these samples by the assay with a probe in exon four-primed were normalised against the Ct values obtained with a probe in exon four, and the results are shown in Figure 18. Interestingly, the normal tissue samples consistently show negative relative expression values, and only two of 105 colorectal cancer tissue samples mix with the normal samples. Hence, setting a threshold at the highest value in the normal samples yields a sensitivity of 98 % for this transcript variant. All the cell lines, and the great majority of the CRC tissue samples (97), show positive relative expression values.

SLC39A14 real-time RT-PCR data in a series of clinical CRC and normal tissue samples, in addition to six colon sell lines, showing the relative expression between the two mutually exclusive splicing variants. The Ct values obtained in the assay with a probe in exon four-primed ("normal exon") is normalised against the Ct values obtained in the assay with a probe in exon four ("cancer exon"), and this gives the relative expression along the y-axis (log-2) for each of the samples (x- axis). It is worth mentioning that all samples that did not cross the Ct line before cycle 34 were considered not to express the specific transcript, and were given the value 34 prior to the calculations. The primers (F and R) and probes (P) for real time RT-PCR:

SLC39A14 ex3_F_TM F GGCCAAGCGCTGTTGAAG (SEQ ID NO: 140) SLC39A14_ex5_R_TM R TCTTCCAGAGGGTTGAAACCAA (SEQ ID NO: 141) SLC39A14_ex4'_P P CTCACTGATTAACCTGGCC (SEQ ID NO: 142)

SLC39A14_ex4_P P ACCGTCATCTCCCTCTG (SEQ ID NO: 143)

The sequences of the transcripts of SLC39A14 (fig 16B) are found in Ensembl release 56 as:

Exon 4':

This exon has Ensembl-id ENSE00001401146, and is no. 4 in the Ensembl- transcript ENST00000359741 (alias SLC39A14-001).

The exon has start-position 22,267,459 and end-position 22,267,628 bases from p- telomer on chromosome 8.

Exon 4:

This exon has Ensembl-id ENSE00000683833, and is no. 4 in the Ensembl- transcripts ENST00000381237, ENST00000240095, and ENST00000289952 (alias SLC39A14-002 (transcript variant 1), SLC39A14-003 (transcript variant 2) and SLC39A14-201 (transcript variant 3)).

The exon has start-position 22,269,550 and end-position 22,269,719 bases from p- telomer on chromosome 8.

Example 4

Materials and methods

Colorectal cell lines and tissue samples.

The project involved analyses of twenty colon carcinoma cell lines from which RNA was isolated by Trizol (Invitrogen, Carlsbad, California, USA). 136 colorectal carcinoma samples, 6 colorectal adenomas (polyps) and 44 normal colonic mucosa samples from cancer patients were also included, from which RNA was isolated by the All prep DNA/RNA mini kit (Qiagen) and the Ribopure™ kit (Applied

Biosystems/Ambion, Foster City, California, USA). Additional RNA samples were included from 14 leukaemia cell lines, 5 embryonal carcinoma cell lines, 2 embryonic stem cells, and 19 miscellaneous healthy organs (Ambion).

Publicly available databases

Sequence information about genes and their different transcripts have been investigated using the EnsembI genome browser and all herein described sequences are in compliance with EnsembI v 60, released November 2010.

Exon microarray analysis

RNA from 99 CRC and 10 normal colonic mucosa samples were analysed by the exon microarrays. Raw data were imported into the XRAY software (version 2.81; Biotique Systems Inc., Reno, Nevada, USA) where quantile normalisation and calculation of probeset expression values were performed and summarized. Only "core" probesets (RefSeq and full-length GenBank mRNAs) were analysed and the expression score for a probeset was defined to be the median of its probe expression scores. For each probeset the log2-ratio of expression level in test samples to that observed in control samples were calculated.

Real time RT-PCT

The cDNA synthesis was performed using the same kit as previously described. The pre-designed commercial quantitative RT-PCR assay was carried out in a fast optical 96-well reaction plate (Applied Biosystems), and the custom-designed assays were performed in standard 96- or 384-well optical reaction plates (Applied Biosystems). Different TaqMan master mixes, reaction volumes, and thermal cycling conditions were used with regard to whether the reactions should be carried out in fast or standard, or 384- or 96-well plates. The TaqMan Fast Universal PCR Master Mix (No AmpErase UNG, Applied Biosystems) was utilized in fast reactions, and the TaqMan Universal PCR Master Mix (AmpErase UNG, Applied Biosystems) was used in standard reactions. The final concentrations of master mix, forward and reverse primers, and probe in the standard reactions were 1 x, 0.9 μΜ of each, and 0.2 μΜ, respectively. In the fast reaction, the end concentrations of master mix and TaqMan Gene Expression Assay (primers and probe) were both 1 x. A total reaction volume of 20 μΙ was used when the reactions were performed in 384- and fast 96-well plates, as distinct from standard 96-well plates, where the total volume per reaction was set to 25 μΙ. RNase free water (Sigma-Aldrich) was added to the correct total volume. In each setup, the amount of starting material (cDNA) was 10 ng.

Multiplex real-time RT-PCR was done as well, adding one primer set, and two different probes with distinct dyes (FAM and VIC), in the same well. In these cases, the concentration of each of the primers where doubled compared to a standard setup, whereas the remainder (plate, master mix, final concentrations, volume, and thermal cycling conditions) was the same as for a standard assay.

Thermal cycling conditions real-time RT-PCR.

Standard Fast

UNG activation : 50 °C 2 min

Polymerase activation : 95 °C 10 min 20 sec

Denaturation : 95 °C 15 sec 1 sec

Annealing and extension : 60 °C 1 min 20 sec

Number of cycles: 40 40

A standard curve was produced by serially diluted universal human reference (UHR) cDNA, synthesised from UHR RNA (Stratagene), of known concentrations. All samples were run in triplicates, and the endogenous control gene assay ACTB (Applied Biosystems) was performed on all the samples. Discussion

SLC39A14, also known as Zrt- and Irt-like protein 14 (ZIP14), is transcribed from the plus strand of cytogenetic band 8p21.3. The Ensembl Genome Browser, release 60, has annotated 12 transcript variants of SLC39A14.

The exon microarray data are shown in Figure 19A. From the exon-wise plot it seems that when exon 4A is expressed, the exon 4B is not expressed, and vice versa, which implies that 4A and 4B are mutually exclusive. This assumption is reinforced by the fact that in known transcripts from this gene, the exons 4A and 4B never exist in the same transcript. As exons 4A and 4B have identical lengths, it is impossible to differentiate between the two splicing events by standard RT-PCR with primers in flanking constitutive exons. Consequently, two real-time RT-PCR assays with exon four specific probes were designed to validate the structure and quantities of the transcripts resulting from these mutually exclusive splicing events (Figure 19B).

In the first-line validation, these assays were performed on 19 of the same samples as examined by the exon microarray analysis, three normals (which were selected based on the exon-wise plots) and 16 CRC tissue samples, in addition to six colon cancer cell lines (Figure 20). This method revealed a clear difference between normal colonic mucosa and CRC samples (both tissues and cell lines). The real-time assay with a probe in exon 4A detected accumulation of PCR product in all three normal mucosa samples (all samples are run in triplicates), but only in two of 16 tumour samples and in none of the six cell lines. Furthermore, the real-time assay with a probe in exon 4B detected PCR product in all cell lines, in addition to 15 of 16 tumour samples, but in none of the normal tissue samples.

Subsequently, these transcripts were further investigated by expanding the sample series of clinical CRC and normal tissue samples. Ct values for exons 4B and 4A were related to each other for each of the assayed samples, and the results are shown in Figure 21. Interestingly, the normal colonic mucosa samples consistently show negative relative expression values (4A with higher expression than 4B; i.e. 4A having the lowest Ct value), and only 8 of 136 colorectal cancer tissue samples are on the negative side. All the CRC cell lines, and the great majority of the CRC tissue samples (128 of 136), show positive 4B vs. 4A relative expression values.

The gene has been further investigated by high throughput RNA-sequencing (Figure 22). Here, we verified the specific inclusion of exon 4B in colorectal cancer cell lines, and of exon 4A in normal colonic mucosa samples. In cancer samples, the two exon 4A and 4B were both present, probably due to the presence of non-cancerous cells within the tumour specimen.

The primers (F and R) and probes (P) for real time RT-PCR:

SLC39A14_ex4_P P ACCGTCATCTCCCTCTG (SEQ ID NO: 143)

The sequences of the transcripts of SLC39A14 (fig 19B) are found in Ensembl release 60 as:

Exon 4A:

This exon has Ensembl-id ENSE00001401146, and is no. 4 in the Ensembl- transcript ENST00000359741 (alias SLC39A14-001). The sequence of this exon has SEQ ID NO: 144.

Exon 4B:

This exon has Ensembl-id ENSE00000683833, and is no. 4 in the Ensembl- transcripts ENST00000381237, ENST00000240095, and ENST00000289952 (alias SLC39A14-002 (transcript variant 1), SLC39A14-003 (transcript variant 2) and SLC39A14-004 (transcript variant 3)).

The exon has start-position 22,269,550 and end-position 22,269,719 bases from p- telomer on chromosome 8. The sequence of this exon has SEQ ID NO: 136 and is in the present application called exon 4 or 4B. Tables

Table 1. Cycling conditions for 5'-RACE.

Table- Appendix-I. Primers 5 cycles 5^cycles 25 cycles

72°C 7 min

Gene Name Type Length Sequence Tm (°C) GC (%)

ABL1 ABL1_ex2_rev Reverse 20 ACCCTGAGGCTCAAAGTCAG 59.5 55

ABL1 ABL1_ex3_rev Reverse 23 TTCCCCATTGTGATTATAGCCTA 640 39

BCR BCR_ex1_forw Forward 20 CAACAGTCCTTCGACAGCAG 59.6 55

BCR BCR_ex13_forw Forward 21 CAGATGCTGACCAACTCGTGT 640 52

BIRC5 BIRC5-6'FAM-R Reverse 20 TCTCCGCAGTTTCCTCAAAT 59.8 45

BIRC5 BIRC5-EX2-L-F Forward 19 GAGGCTGGCTTCATCCACT 60. 58

BIRC5 BIRC5-EX1- F Forward 20 AGAACTGGCCCTTCTTGGAG 60.8 55

BIRC5 BIRC5-EX2-K-F Forward 20 GCCCAGTGTTTCTTCTGCTT 59.5 50

BIRC5 BIRC5_ex _Rev Reverse 20 TCTCCGCAGTTTCCTCAAAT 59.8 45

C4BPB C BPB_ex1_F Forward 26 CCTTGCTGGGAAGCCCTAACTCTGGA 71.7 58

C4BPB C BPB_ex2_R Reverse 25 ACGCAACCATAAGACAGCACGCACA 70.6 52

C4BPB C BPB_ex2_nest_R Reverse 25 GGCTGGAATTCACCCAGCTCAGACA 70.5 56

CST1 CST1_ex1_F Forward 25 TGCGGGTACTAAGAGCCAGGCAACA 70.9 56

CST1 CST1_ex3_R Reverse 2 CGAATGGCCTGGCACAGATCCCTA 71.0 58

CST1 CST1_ex3_nest_R Reverse 27 TGACACCTGGATTTCACCAGGGACCTT 71.7 52

ETV6 ETV6_ex5_forw Forward 20 CACTCCGTGGATTTCAAACA 59.5 45

FZD10 FZD10_ex1_F Forward 25 TTTATGCTGCTGGTGGTGGGGATCA 71.3 52

FZD10 FZD10_ex1_R Reverse 25 CCGTGGTGAGTTTTCTGGGGATGCT 71.3 56

FZD10 FZD10_ex1_nest_R Reverse 25 GCCGCCAGGATCTTCCAGTAATCCA 71.3 56

GJB6 GJB6_ex3_F Forward 25 TTCGGATAGAGGGGTCGCTGTGGTG 72.1 60

GJB6 GJB6_ex3_R Reverse 25 GCAGCATGCAAATCACAGACGCAGA 71.2 52

GJB6 GJB6_ex3_nest_R Reverse 25 AACAAGGTTGGGGCAGGGGTCAATC 72.0 56

GPR177 GPR177_ex2_R Reverse 20 GGAGGGGAATGTGAACAGAA 57.0 50

GPR177 GPR177_ex1_F Forward 20 TCTGCTCGTGTTCCAAATCA 57.1 45

HOXB13 HOXB13_ex1_F Forward 25 CAGCCAGATGTGTTGCCAGGGAGAA 71.2 56

HOXB13 HOXB13_ex2_R Reverse 25 CTTGCGCCTCTTGTCCTTGGTGATG 70.9 56

HOXB13 HOXB13_ex2_R_alt2 Reverse 28 TAAGGGGTAGCGCTGTTCTTCACCTTGG 72.5 54

HOXC11 HOXC11_ex1_F Forward 25 ACAAATCCCAGCTCGTCCGGTTCAG 71.4 56

HOXC11 HOXC11_ex2_R Reverse 25 CCCTGGCCACAGTCCAGTTTTCCAC 71.6 60

HOXC11 HOXC11_ex2_nest_R Reverse 25 CCGGTCTGCAGGTTACAGCAGAGGA 70.6 60

Hs.446400 Hs. 6 00_F Forward 20 CAGAGCTGCATCCTTATGGT 55.1 50

Hs.446400 Hs. 6400_R Reverse 20 AGCTGCAAGTTGTTGTTCCA 56.5 45

MIER1 MIER1_ex9_F Forward 22 CCATCAGAAGACTGGAAAAAGG 58.3 45

MIER1 MIER1_ex10_R Reverse 22 TGCTTCTACACCCTTCTCATCA 57.5 45

MTHFD2L MTHFD2L_ex5_F Forward 20 GACCCAAGAGTCAGCGGTAT 56.5 55

MTHFD2L MTHFD2L_ex7_R Reverse 20 G ATCTTCCAG CCACAACCAC 57.4 55

NKAIN2 NKAIN2_ex8_F Forward 27 TGGCTATCAAGGGCCTCAGAAGACATC 70.0 52

NKAIN2 NKAIN2_ex10_R Reverse 25 CAGGAAATCCAAGATGGGCGTGTCC 71.5 56

NKAIN2 NKAIN2_ex10_nest_R Reverse 25 CAAGTGGAATTGGTGTGTGCGTGCT 70.0 52

PBX1 PBX1_ex3_rev Reverse 21 TGCTCCACTGAGTTGTCTGAA 59.1 48

PBX1 PBX1_ex5_rev Reverse 20 GGGTTGCTGAGATGGGAATA 59.9 50

PRRX1 PRRX1_ex _F Forward 25 TAGACCTGGAGGAAGCCGGGGACAT 71.9 60

PRRX1 PRRX1_ex _R Reverse 25 TAATCGGTGGGTCTCGGAGCAGGAC 71.3 60

PRRX1 PRRX1_ex _nest_R Reverse 25 GTGTCCGCTCAAAGACACGCTCCAA 71.4 56

PRRX1 PRRX1_int1_ex3_R Reverse 25 CCCAGCTTTGGTGGCACTTCTGTGA 71.3 56

PRRX1 PRRX1_int1_ex _R Reverse 28 TCAGGGAAAACGTGAAACTCCTCTTGTC 69.2 46

PRRX2 PRRX2_ex3_F Forward 25 GCCCACCGCCCTGAGTCCAGATTAT 72.2 60

PRRX2 PRRX2_ex _R Reverse 25 AGGTCCTTGGCAGGCTCTTCCACCT 71.4 60

PRRX2 PRRX2_ex _nest_R Reverse 25 CAAGGGTTGTGGGCTGCAGTCTCTG 71.0 60

RAD51L1 RAD51 L1_ex5_F Forward 27 CCCACCAACATGGGAGGATTAGAAGGA 71.2 52

RAD51L1 RAD51 L1_ex8 _R Reverse 25 AGCTGGAGACACCAGGTCTGCCTGA 70.3 60

RAD51L1 RAD51 L1_ex8_nest_R Reverse 25 CTGAGAAGCCAGGGCTCCACTCAGA 70.0 60

RUNX1 RUNX1_ex2_rev Reverse 20 CGTGGACGTCTCTAGAAGGA 58.0 55

SERPINB7 SERPINB7_ex3_F Forward 25 TTGGGCGCTCAAGATGACTCCCTCT 71.2 56

SERPINB7 SERPINB7_ex5_R Reverse 25 GTCAACTCGCTCCACTTTGGCATCG 70.9 56

SERPINB7 SERPINB7_ex6_R Reverse 26 GAAGGCTGATTGCCACTTGCCTTTGA 71.5 50

TCF3 TCF3_ex15_forw Forward 19 CACCCTCCCTGACCTGTCT 60.0 63

TCF3 TCF3_ex17_forw Forward 20 GTGACATCAACGAGGCCTTT 60.1 50

TFPT TFPT_ex3_F Forward 25 CACATCCTGGAGAGCGAGCTGGAGA 70.8 60

TFPT TFPT_ex _R Reverse 25 TCCTGCTGCAGCCTCCGAGTTATCC 71.7 60

TFPT TFPT_ex _nest_R Reverse 2 CCTGTTCAGGACCCGCTCGTTCAC 70.8 63

TFR2 TFR2_ex7_F Forward 25 TCAGGACTTCGGGGCTCAAGGAGTG 71.6 60

TFR2 TFR2_ex8_R Reverse 25 GCTGGGAAGGCCTGATGATGCAACT 71.5 56

TFR2 TFR2_ex8_nest_R Reverse 25 TGTAGGGGTCTCCAGTTCCCAGGTG 69.4 60 Universal NUP Forward 23 AAGCAGTGGTATCAACGCAGAGT 60.8 48

Universal UPM - Long Forward 45 CTAATACGACTCACTATAGGGCAAGCAGT 80. 47

GGTATCAACGCAGAGT

Universal UPM - Short Forward 22 CTAATACGACTCACTATAGGGC 51.3 45

USP11 USP11_eks6_R Reverse 18 G CCTGG CTG ACCCTTG A A 58.8 61

USP11 USP11_ex5_F Forward 18 GAGCGGTTTCTGGTGGAG 55.7 61

VNN1 VNN1_ex5_F Forward 25 TGCACACTGTGGAAGGGCGCTATTA 69.7 52

VNN1 VNN1_ex7_R Reverse 27 G GCTTCAG ACTAAACAAG CG TCCG TCA 70.8 52

VNN1 VNN1_ex6_nest_R Reverse 25 CTG GGTTCCG AAAG TG CCACTG AG G 71.8 60

WIF1 WIF1_ex9_F Forward 26 GAACCTGCCATGAACCCAACAAATGC 71.4 50

WIF1 WIF1_ex10_R Reverse 25 GCCGCTCCTCGGCCTTTTTAAGTGA 72.5 56

WIF1 WIF1_ex9_nest_R Reverse 25 ATGGCAGGTTCCATGTGCACCACAG 71.7 56

ZDHHC20 ZDHHC20_ex1_F Forward 18 CTGGAGCGTCCGAGTCAC 56.3 67

ZDHHC20 ZDHHC20 ex2 R Reverse 22 CAACGGTCTTTCCATTTTCTTC 58.5 41

Tables- Appendix-II

Abbreviations:

T = Primary tumour sample

C = Cell line

N = Number of times sequenced

Table-A-II- 1. Exon positions from RAD51L1.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000182185; transcribed from plus strand; start position : 67,356,262 bp from chromosome 14 p-telomere; EnsembI release 50).

Table-A-II- 2. Exon positions from NKAIN2.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000188580; transcribed from plus strand; start position : 124,166,985 bp from chromosome 6 p-telomere; Ensembl release 50).

Table- A-II- 3. Exon positions from VNN1.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000112299; transcribed from minus strand; start position : 133,076,881 bp from chromosome 6 p-telomere; Ensembl release 50).

Table-A-II- 4. Exon positions from C4BPB.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000123843; transcribed from plus strand; start position : 205,328,835 bp from chromosome 1 p-telomere; Ensembl release 50).

^*ENST00000391923 lacks base pairs 496-614

Table- A-II- 5. Exon positions from HOXC11.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000123388; transcribed from plus strand; start position : 52,653,177 bp from chromosome 12 p-telomere; Ensembl release 50).

Table-A-II- 6. Exon positions from TFR2.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000106327; transcribed from minus strand; start position : 100,077,109 bp from chromosome 7 p-telomere; Ensembl release 50).

^* One clone lacks base pairs 8551 -8714

" One clone lacks base pairs 9470-9570

Table-A-II- 7. Exon positions from SERPINB7.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000166396; transcribed from plus strand; start position : 59,571,257 bp from chromosome 18 p-telomere; EnsembI release 50).

Table-A-II- 8. Exon positions from TFPT.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000105619; transcribed from minus strand; start position : 59,310,867 bp from chromosome 19 p-telomere; Ensembl release 50).

^*ENST00000339150 lacks base pairs 163-268

Table-A-II- 9. Exon positions from GJB6.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000121742; transcribed from minus strand; start position : 19,704,456 bp from chromosome 13 p-telomere; Ensembl release 50).

Table-A-II- 10. Exon positions from PRRX1.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000116132; transcribed from plus strand; start position : 168,899,937 bp from chromosome 1 p-telomere; Ensembl release 50).

*Exon δ in sequence PRXX1 B and PRRX1 C is a retention of the intron between exons

** The exon lacks bases 53,625-53,778

Table- A-II- 11. Exon positions from PRRX2.

Start/end sequence positions are indicated relative to start of exon 1

(ENSG00000167157; transcribed from plus strand; start position : 131,467,741 bp from chromosome 9 p-telomere; Ensembl release 50).

Claims

1. A method for the detection of abnormal gene expression of SLC39A14, said method comprising a) identifying an expression level of at least one RNA transcript variant of SLC39A14 in a sample obtained from a test subject, b) comparing the expression level of said at least one RNA transcript variant of SLC39A14 with a reference obtained from a reference subject, c) selecting a desired sensitivity, d) selecting a desired specificity, e) indicating the test subject as likely to have an abnormal gene expression, if the expression level of said at least one RNA transcript variant of SLC39A14 in the sample obtained from a test subject is significantly different from the reference, and indicating the test subject as unlikely to have an abnormal gene expression, if the expression level of said at least one RNA transcript variant SLC39A14 is equal to the reference.

2. A method for the detection of abnormal gene expression of SLC39A14, said method comprising the step of a) determining an expression level of at least one RNA transcript variant SLC39A14 in a sample obtained from a test subject.

3. The method according to claim 2 further comprising the steps of b) comparing the expression level of said at least one RNA transcript variant of SLC39A14 with a reference obtained from a reference subject, c) selecting a desired sensitivity, d) selecting a desired specificity, e) indicating the test subject as likely to have an abnormal gene expression, if the expression level of said at least one RNA transcript variant of SLC39A14 in the sample obtained from a test subject is significantly different from the reference, and indicating the test subject as unlikely to have an abnormal gene expression, if the expression level of said at least one RNA transcript variant of

SLC39A14 is equal to the reference.

4. The method according to claims 1-3, wherein the expression level of said at least one RNA transcript variant of SLC39A14 in the test subject is higher than the reference subject.

5. The method according to claims 1-4, wherein said sample is selected from the group consisting of blood, serum, plasma, faeces, tissue biopsy, culture cells.

6. The method according to claims 1-5, wherein the abnormal expression pattern is indicative of cancer or a precursor to cancer or an inflammatory disease or a viral infection or a metabolic disease in the test subject.

7. The method according to claim 6, wherein the cancer is colorectal cancer or the precursor to cancer is colorectal adenomas or dysplastic cells.

8. The method according to claims 1-7, wherein the SLC39A14 transcript variant is selected from the group consisting of transcript 1 (SEQ ID NO: 137), transcript 2 (SEQ ID NO: 138), or transcript 3 (SEQ ID NO: 139).

9. The method according to claims 1-8, wherein the SLC39A14 exon 4B is ENSE00000683833 (SEQ ID NO: 136).

10. The method according to claims 1-9, wherein exon 4B (SEQ ID NO: 136) is present and exon 4A (SEQ ID NO: 144) is not present in the sample from the test subject.

11. Use of at least one RNA transcript variant selected from the list consisting of (SEQ ID NO: 136), (SEQ ID NO: 137), (SEQ ID NO: 138), and (SEQ ID NO: 139) as a biomarker.

12. The use according to claim 11, wherein said biomarker is a biomarker for diagnosing, prognosing, monitoring and/or treatment selection for a cancer or a precursor to cancer.

13. The use according to claim 12, wherein said cancer is colorectal cancer.