Poster: Genomics Resources

Abs # 950: High throughput cloning and analysis of Arabidopsis genes predicted by computer algorithms and comparative genomics

Presenter:	Xiao, Yongli , yxiao@tigr.org
Authors	Xiao, Yongli  (A)   Smith, Shannon  (A)   Ishmael, Nadeeza  (A)   Kumar, Nikhil  (A)   Redman, Julia  (A)   Ayele, Mulu  (A)   Haas, Brian J (A)   Town, Christopher D (A)
Affiliations:	(A): The Institute for Genomic Research

We are interested in elucidating the structure and function of Arabidopsis genes that have no known homologs outside Arabidopsis. A large number of these are the so-called “hypothetical genes” that are based only upon computer prediction, requiring two concordant predictions to be included in the current TIGR genome annotation. Recently, we have also used comparative genomic analysis of ~400,000 whole genome shotgun reads from Brassica to infer the existence of additional genes that have neither good computational nor database support. Our analysis indicates the presence of ~10,000 regions of sequence conservation in the parts of the Arabidopsis genome not presently annotated, which could suggest many novel genes without reliable gene prediction. At TIGR, we have developed a high throughput method for the cloning and analysis these predicted genes. To test their expression, we used cDNA populations from a number of plant tissues and treatments. Of the 893 hypothetical genes tested, 596 are expressed in our cDNA mixture. Of the first set of 133 genes predicted by comparative genomics, 57 are expressed. The full-length cDNA sequences of these predicted genes were obtained by sequencing and assembling their 5’ and 3’ RACE products. So far, we have obtained a total 506 full-length or partial cDNA sequences of Arabidopsis hypothetical genes and 50 genes predicted by Brassica sequence conservation. The cDNA sequences from 212 hypothetical genes display differences from their predicted gene structures. In total, 83 hypothetical genes have alternatively spliced transcripts and 103 genes display more than one polyadenylation site. Thus many of both the hypothetical and Brassica-predicted genes are expressed. Functional analysis of these genes is in progress.