TY - JOUR
T1 - Non-EST-based prediction of novel alternatively spliced cassette exons with cell signaling function in Caenorhabditis elegans and human
AU - Leparc, German Gaston
AU - Mitra, Robi David
N1 - Funding Information:
We would like to thank Gary Stormo, Justin Sonnenburg, Katherine Varley and Jason Gertz for reviewing our manuscript. Tim Schedl and Elaine Mardis for their excellent suggestions concerning our experimental approach. Total RNA samples of C. elegans were graciously donated by Jennifer Davila-Aponte from the Sean Eddy lab. We would also like to thank Yun Yue, Lee Tessler and Michael Brooks for helpful discussions. This work and the Open Access publication charges were funded by NIH grant no. 5P50HG003170-03 and GATP training grant no. T32-HG000045, Whitaker Foundation (St. Louis, MO).
PY - 2007/5
Y1 - 2007/5
N2 - To better understand the complex role that alternative splicing plays in intracellular signaling, it is important to catalog the numerous splice variants involved in signal transduction. Therefore, we developed PASE (Prediction of Alternative Signaling Exons), a computational tool to identify novel alternative cassette exons that code for kinase phosphorylation or signaling protein-binding sites. We first applied PASE to the Caenorhabditis elegans genome. In this organism, our algorithm had an overall specificity of ≥76.4%, including 33 novel cassette exons that we experimentally verified. We then used PASE to analyze the human genome and made 804 predictions, of which 308 were found as alternative exons in the transcript database. We experimentally tested 384 of the remaining unobserved predictions and discovered 26 novel human exons for a total specificity of ≥41.5% in human. By using a test set of known alternatively spliced signaling exons, we determined that the sensitivity of PASE is ∼70%. GO term analysis revealed that our exon predictions were found in the introns of known signal transduction genes more often than expected by chance, indicating PASE enriches for splice variants that function in signaling pathways. Overall, PASE was able to uncover 59 novel alternative cassette exons in C. elegans and humans through a genome-wide ab initio prediction method that enriches for exons involved in signaling.
AB - To better understand the complex role that alternative splicing plays in intracellular signaling, it is important to catalog the numerous splice variants involved in signal transduction. Therefore, we developed PASE (Prediction of Alternative Signaling Exons), a computational tool to identify novel alternative cassette exons that code for kinase phosphorylation or signaling protein-binding sites. We first applied PASE to the Caenorhabditis elegans genome. In this organism, our algorithm had an overall specificity of ≥76.4%, including 33 novel cassette exons that we experimentally verified. We then used PASE to analyze the human genome and made 804 predictions, of which 308 were found as alternative exons in the transcript database. We experimentally tested 384 of the remaining unobserved predictions and discovered 26 novel human exons for a total specificity of ≥41.5% in human. By using a test set of known alternatively spliced signaling exons, we determined that the sensitivity of PASE is ∼70%. GO term analysis revealed that our exon predictions were found in the introns of known signal transduction genes more often than expected by chance, indicating PASE enriches for splice variants that function in signaling pathways. Overall, PASE was able to uncover 59 novel alternative cassette exons in C. elegans and humans through a genome-wide ab initio prediction method that enriches for exons involved in signaling.
UR - http://www.scopus.com/inward/record.url?scp=34250872207&partnerID=8YFLogxK
U2 - 10.1093/nar/gkm187
DO - 10.1093/nar/gkm187
M3 - Article
C2 - 17452356
AN - SCOPUS:34250872207
SN - 0305-1048
VL - 35
SP - 3192
EP - 3202
JO - Nucleic acids research
JF - Nucleic acids research
IS - 10
ER -