Author:

Keywords:

Alternative Splicing, CANCER

Abstract:

Alternative splicing (AS) is the process by which multiple mature mRNA sequences can be generated from the same precursor mRNA (pre-mRNA) upon the differential joining of exonic sequences limited by 5' and 3' splice sites. These variants of mRNA can be produced from precursor mRNA by extending, shortening, skipping or including, and retaining intron sequences. Despite specific roles in tissues, stages of development and cell compartments etc., these isoforms have been related to alteration of splicing machinery causing disease (including cancer) and hence, targeted for therapeutic and diagnostic tools. We have developed an algorithm to analyze AS events and addressed many previously encountered problems with the help of filters, robust summarization algorithms and general linear models. We have applied these algorithms to analyze the data of non-small-cell-lung- carcinoma using Human Exon Junction Arrays (HJAY, provided by Affymetrix) for known Alternative splicing events. To ease the validation process, we have focused only in skipped exons (cassette events). These algorithms combine different statistical techniques that jointly analyze the different probes that interrogate the event with the aid of the Irwin-Hall Distribution. The final result of this analysis is a statistical p-value that is used to rank the exons as likely candidates of AS events. In addition, we have developed a program to automatically generate HTML reports for given AS events. 100% of top-ranked AS events have been validated in wet lab. This result is especially interesting if compared with the use of previous algorithms on earlier arrays (validation rate less than 30%)