Browse Theses

Novel methods are used to study the extent and regulation of alternative splicing in the human and mouse transcriptomes. Using splicing graphs and whole genome alignments, the largest set to date of orthologous alternative splicing events is discovered, including over 2,162 cassette exons present in both the human and mouse transcriptomes. These high confidence cassette exons, alternative 3' and alternative 5', sequences are found to be flanked by unusually conserved intronic sequences. The alternative portions of the exons themselves is also found to be more conserved than in constitutive exons. A new-class of very small 3bp alternative 3' splicing events is found to be quite prevalent as well. It is not clear if these events are the result of noise in the spliceosome or are functional biologically.

Using a novel algorithm called ExonWalk , full length transcripts that are implied by the fragmentary EST and mRNA data are created. This approach is shown to be very competitive with both current gene building and gene finding approaches. ExonWalk reconstructs the coding exons reviewed RefSeq genes that are spliced with 87% sensitivity and 83% specificity and the individual base pairs are found with 95% sensitivity and 86% specificity. Using orthologous splicing as an additional filter to transcript counts and assembling exons rather than entire transcripts helps to remove much of the noise present in the EST and mRNA Genbank.

Using the splicing graphs to model gene structure splicing, specific GeneChips are investigated for 23 normal mouse tissues. Numerous alternative splicing events are shown to occur in these tissues. Cassette exons that are preferentially included or excluded in brain tissues relative to non-brain tissues are discovered and shown to be flanked by very highly conserved sequences and to have differential levels of known RNA binding motifs. Paralogous MAGUK genes with similar alternative splicing profiles are identified and it is shown that although these paralogous exons are still both alternatively spliced, the primary sequence has diverged significantly. This may imply a regulation more complex than originally thought, or possibly regulation via secondary mRNA structure.