FirstEF* (First Exon Finder) is a 5' terminal exon and promoter prediction program. It consists of different discriminant functions structured as a decision tree. The probabilistic models are optimized to find potential first donor sites and CpG-related and non-CpG-related promoter regions based on discriminant analysis. For every potential first donor site (GT) and an upstream promoter region, FirstEF decides whether or not the intermediate region can be a potential first exon, based on a set of quadratic discriminant functions.
FirstEF calculates the a posteriori probabilities of exon, donor, and promoter for a given GT and an upstream window of length 570 bp. If Probability(Exon)>0.5, Probability(Donor)>0.4, and Probability(Promoter)>0.4, FirstEF outputs the region as a predicted exon.
The output contains the following 8 fields:
|No.||Serial Number of the predicted exon cluster (If two predicted exons are separated by less than 1000 bp, they are considered as first-exon predictions of the same gene and ranked based on posterior probabilities). Only high scoring exons within each cluster are presented.|
|Promoter:||Predicted promoter of length 570 bp|
|P(promoter):||A posteriori probability of promoter for a given window of size 570 bp|
|Exon:||Predicted exon boundaries|
|P(exon):||A posteriori probability of exon for a given GT and promoter region|
|P(donor):||A posteriori probability of donor for a given GT|
|CpG Window:||Boundaries of the CpG window of length 201, if the exon is CpG-related. Else, output "Non-CpG-related"|
|Rank:||Rank of the first exon within a cluster|
If you have problems, comments or suggestions, please contact Ramana Davuluri, Ivo Grosse or Michael Zhang
*FirstEF (c) 2001 Cold Spring Harbor Laboratory is available for free to non-profit institutions using it for non-commercial purposes. Commercial users may obtain a site license for a fee. For information about obtaining a commercial or academic user license contact Lorraine McInerny by email, phone: 516-367-8301, or fax: 516-367-8350. ALL RIGHTS RESERVED.
We encourage users interested in genome-wide promoter and first-exon analyses to install a local copy of FirstEF. The web server implementation of FirstEF is intended for quick and local searches of promoters and first exons, and we restrict the maximum input file size to 100 KB in order to discourage users from using the FirstEF web server for genome-wide searches of promoters and first-exons.
The downloadable version of FirstEF is identical to the web server version of FirstEF, but it includes two additional output parsers, firstef_genome_parser.pl and firstef_promoter_parser.pl, which might be useful for genome-wide analyses of promoters and first-exons.
firstef_genome_parser.pl merges first-exon predictions from both strands, and we used that parser to obtain the genome-wide predictions of first exons in the human genome.
firstef_promoter_parser.pl merges promoter predictions from both strands, and we used that parser to compare the promoter-prediction accuracies of PromoterInspector and FirstEF.
REFERENCE: Ramana V. Davuluri, Ivo Grosse, Michael Q. Zhang: Computational identification of promoters and first exons in the human genome. Nature Genetics (2001) 29, 412--417.