SVM-BPfinder (beta)
[A tool for mammalian U2 branch point prediction]
Upload your intron 3' end sequences in fasta format (max. 1000 sequences).
Sequences should NOT:
1) be longer than 500 nt or shorter than 20 nt;
2) contain characters other than a c g t A C G T;

Example fasta file:

Select one from the available species-specific BP models.
You can also limit the output to the AG-dinucleotide exclusion zone (AGEZ) only.
Submit and please wait. (time is directly proportional to input size)

Results are shown in plain text, tab delimited, one line per BP candidate.
Output fields:
seq_id - Sequence Identifier
agez - AG dinucleotide Exclusion Zone length
ss_dist - Distance to 3' splice site
bp_seq - BP sequence (nonamer; from -5 to +3 relative to the BP adenine)
bp_scr - BP sequence score using a variable order Markov model
y_cont - Pyrimidine content between the BP adenine and the 3' splice site
ppt_off - Polypyrimidine tract offset relative to the BP adenine
ppt_len - Polypyrimidine tract length
ppt_scr - Polypyrimidine tract score
svm_scr - Final BP score using the SVM classifier

Upload fasta file:

or paste sequences in fasta format: (overrides file input)

Select organism:
Output results for:
all intron
AGEZ only

3 and GO!

A standalone version of this tool is also available for use (linux only). You can get it here. Please read the README file for details on how to install and use.

Important: In standalone version, input sequence must be in lowercase. Species name uses the 4-letter code: Hsap, Ptro, Mmus, etc.

Genome-Wide Association between Branch Point Properties and Alternative Splicing. Corvelo A, Hallegger M, Smith CW, Eyras E.
PLoS Comput Biol. 2010 Nov 24;6(11):e1001016. PMID: 21124863
Supplementary Data

This tool was developed by Andre Corvelo and Eduardo Eyras. Universitat Pompeu Fabra, Barcelona Spain 2010. It uses SVMlight.