Plant & Animal Genome IX Conference

Poster: Sequencing & EST
P01_44.html

EXPERIMENTAL VALIDATION OF "HYPOTHETICAL PROTEINS" IN ARABIDOPSIS THALIANA

The Institute for Genomic Research, 9712 Medical Center Drive, Rockville, MD 20850, USA

Analysis and annotation of chromosomes 2 and 4 of Arabidopsis thaliana revealed that about one a third of the genes were supported only by computational predictions with no similarity to known proteins or EST sequences. Many of these "hypothetical proteins" are closely related to sequences found elsewhere in the genome. We have begun to construct and analyze protein families containing "hypothetical proteins" with a view to experimental verification of the genes and their structures. A partial proteome consisting of 14,767 proteins was analyzed and clustered into families using the domain-based GEANFAMMER package. We selected 21 families for further analysis and plan to characterize the structure of ~25% representative genes from each family. We will attempt to isolate full-length cDNAs for these hypothetical genes by 3', 5'-RACE using mRNA isolated from a variety of tissues/treatments. The 3', 5'-RACE products will be sequenced and aligned with genomic sequence to generate the correct gene structure. To date, we have isolated RNA from 7 different tissues/treatments and have made primers from 5 different genes representative of 2 families for RT-PCR. In the first phase we will concentrate on "hypothetical proteins" from chromosome 2 which was completely sequenced and annotated at TIGR. Total number of predicted genes (Lin et al, Nature 402:761-768, 1999) was 4037 out of which 1094 were hypothetical. Data released since this time has reduced the number of hypothetical proteins (genes with no GenBank support) to 847. The results from this analysis will be presented.