ALTERNATIVE GENE FORM DISCOVERY AND CANDIDATE GENE SELECTION FROM GENE INDEXING PROJECTS

Several efforts are under way to partition single-read expressed seque nce tag (EST), as well as full-length transcript data, into large-scal e gene indices, where transcripts are in common index classes if and o nly if they share a common progenitor gene. Accurate gene indexing fac ilitates gene expression studies, as well as inexpensive and early gen e sequence discovery through assembly of ESTs that are derived From ge nes that have not been sequenced by classical methods. We extend, corr ect, and enhance the information obtained from index groups by splitti ng index classes into subclasses based on sequence dissimilarity (dive rsity). Two applications of this are highlighted in this report. First it is shown that our method can ameliorate the damage that artifacts, such as chimerism, inflict on index integrity. Additionally, we demon strate how the organization imposed by an effective subpartition can g reatly increase the sensitivity of gene expression studies by accounti ng for the existence and tissue-or pathology-specific regulation of no vel gene isoforms and polymorphisms. We apply our subpartitioning trea tment to the UniGene gene indexing project to measure a marked increas e in information quality and abundance (in terms of assembly length an d insertion/deletion error) after treatment and demonstrate cases wher e new levels of information concerning differential expression of alte rnate gene forms, such as regulated alternative splicing, are discover ed.