Seqping: Gene Prediction Pipeline for Plant Genomes using Self-Trained Gene Models and Transcriptomic Data

ABSTRACT Summary Although various software are available for gene prediction, none of the currently available gene-finders have a universal Hidden Markov Models (HMM) that can perform gene prediction for all organisms equally well in an automatic fashion. Here, we report an automated pipeline that performs gene prediction using selftrained HMM models and transcriptomic data. The program processes the genome and transcriptome sequences of a target species through GlimmerHMM, SNAP, and AUGUSTUS training pipeline that ends with the program MAKER2 combining the predictions from the three models in association with the transcriptomic evidence. The pipeline generates species-specific HMMs and is able to predict genes that are not biased to other model organisms. Our evaluation of the program revealed that it performed better than the use of the closest related HMM from a standalone program. Availability and Implementation Distributed under the GNU license with free download at http://sourceforge.net/projects/seqping and http://genomsawit.mpob.gov.my . Contact [email protected] Supplementary information Supplementary data are available at Bioinformatics online.