The CONTRA (CONditional TRAining for Biosequence Analysis) Project is a collection of programs which apply discriminatively-trained probabilistic models (specifically conditional random fields and their variants) to a diverse array of sequence labeling problems in bioinformatics.

Currently, the CONTRA family includes:

CONTRAlign, a protein sequence aligner
CONTRAfold, an RNA secondary structure folder
CONTRAST, a multiple sequence de novo gene predictor