Algorithms and software for support of gene identification experiments

MOTIVATION: Gene annotation is the final goal of gene predictionalgorithms. However, these algorithms frequently make mistakes andtherefore the use of gene predictions for sequence annotation is hardlypossible. As a result, biologists are forced to conduct time-consuming geneidentification experiments by designing appropriate PCR primers to testcDNA libraries or applying RT-PCR, exon trapping/amplification, or othertechniques. This process frequently amounts to 'guessing' PCR primers ontop of unreliable gene predictions and frequently leads to wasting ofexperimental efforts. RESULTS: The present paper proposes a simple andreliable algorithm for experimental gene identification which bypasses theunreliable gene prediction step. Studies of the performance of thealgorithm on a sample of human genes indicate that an experimental protocolbased on the algorithm's predictions achieves an accurate geneidentification with relatively few PCR primers. Predictions of PCR primersmay be used for exon amplification in preliminary mutation analysis duringan attempt to identify a gene responsible for a disease. We propose asimple approach to find a short region from a genomic sequence that withhigh probability overlaps with some exon of the gene. The algorithm isenhanced to find one or more segments that are probably contained in thetranslated region of the gene and can be used as PCR primers to selectappropriate clones in cDNA libraries by selective amplification. Thealgorithm is further extended to locate a set of PCR primers that uniformlycover all translated regions and can be used for RT-PCR and furthersequencing of (unknown) mRNA.