GenomeThreader Gene Prediction Software

GenomeThreader is a software tool to compute gene structure predictions. The gene structure predictions are calculated using a similarity-based approach where additional cDNA/EST and/or protein sequences are used to predict gene structures via spliced alignments. GenomeThreader was motivated by disabling limitations in GeneSeqer, a popular gene prediction program which is widely used for plant genome annotation.

Features

• Intron Cutout Technique:
  The intron cutout technique allows to overcome the time and space limitations of the dynamic programming (DP) algorithms used in GeneSeqer, in particular, when applied to organisms containing long introns.
• Baysian Splice Site Models (BSSMs):
  With BSSMs it is possible to assign probabilities to GT donor, GC donor, and AG acceptor sites. This information is used in the DP to get the exact exon/intron boundaries right.
• Combination of cDNA/EST Based Spliced Alignments with Protein Based Spliced Alignments:
  After (spliced) aligning the supplied cDNAs/ESTs and protein sequences onto the genomic template, GenomeThreader computes consensus spliced alignments. Consensus spliced alignments combine several spliced alignments to resolve the complete gene structure and to uncover alternative splicing.
• Incremental Updates:
  When the used cDNA/EST or protein database is updated, a common approach was to redo the complete mapping. With GenomeThreader, you can combine newly computed spliced alignments with precomputed spliced alignments to quickly recompute consensus spliced alignments.
• XML:
  The additional GenomeThreader XML output conforms to our gthXML standard GenomeThreader.rng.txt. With the included script XML2GFF.py, it is possible to convert gthXML output to the GFF format. A variety of gthXML-specific tools can be found here.
• gthDB:
  We also provide a schema and load script for gthDB, which permits storage and query of GenomeThreader output in a relational format.

References have been omitted for brevity; you can find them and more details on the implementation in the GenomeThreader paper. How to take advantage of these features and many more is described in depth in the GenomeThreader manual. All mentioned files and scripts are also part of the GenomeThreader distribution (see below).

Availability

GenomeThreader is available free of charge for non-commercial research institutions. To obtain a copy, please fill out the license agreement and send it to us by fax or email (as described in the document). As soon as we receive your license agreement, we send you a username and password by email which you can use to download GenomeThreader.

Examples

• Evaluation cases described in Gremme et al. 2005 (see below)
• A 16.6Kb rice gene structure tractable with GenomeThreader (using both an intron cutout technique and without), but beyond GeneSeqer's limitations.
• A 125Kb intron-containing human gene structure.
• Small samples of gzip'ed plain text and XML GenomeThreader output.

Users

The following sites use GenomeThreader. This list is not intended to be comprehensive.

• MIPS (Munich Information Center for Protein Sequences), IBI at GSF Neuherberg (for plant genome annotation)
• University of Freiburg, Plant Biotechnology (to annotate Physcomitrella patens)
• PlantGDB
• Waksman Institute, Rutgers University

If you want to appear on this list, please drop me a note.

Developers

GenomeThreader is being actively developed by the following individuals:

• Gordon Gremme (primary developer)
• Stefan Kurtz (Vmatch, Mkvtree, libkurtz)
• Volker Brendel (BSSMs, conceptual ideas)
• Michael E Sparks (BSSMs, gthXML, gthDB)

Publications

Please cite the following article in publications about research using GenomeThreader:

• G. Gremme, V. Brendel, M.E. Sparks, and S. Kurtz. Engineering a software tool for gene structure prediction in higher organisms. Information and Software Technology, 47(15):965-978, 2005