nGASP--the nematode genome annotation assessment project

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Limited genomic annotation exists for Caenorhabditis species beyond C. elegans.
• The nematode genome annotation assessment project (nGASP) was initiated to evaluate gene prediction software accuracy.
• Seventeen international groups participated, submitting 47 prediction sets for C. elegans.

Purpose of the Study:

• To objectively assess protein-coding gene prediction software accuracy in C. elegans.
• To apply this knowledge to annotate genomes of four additional Caenorhabditis species and other nematodes.
• To establish a baseline for gene prediction accuracy in Caenorhabditis genomes.

Main Methods:

• Comparative analysis of 47 gene prediction sets against curated reference gene models from WormBase.
• Evaluation of various gene-finding approaches, including transcript-, protein-, and multi-genome alignments, and ab initio methods.
• Assessment of gene prediction performance based on sensitivity and specificity metrics.

Main Results:

• 'Combiner' algorithms, integrating multiple alignment types and predictions, achieved the highest accuracy (78% median sensitivity, 42% specificity).
• Gene finders utilizing transcript/protein alignments ranked second, followed by multi-genome and ab initio methods.
• Genes with unusual exon content, splice sites, or poorly conserved orthologs presented challenges for prediction software.

Conclusions:

• nGASP established a benchmark for gene prediction accuracy in nematode genomes.
• The study guided the selection of optimal gene-finders for annotating newly sequenced nematode genomes.
• New gene sets were generated for C. briggsae, C. remanei, C. brenneri, C. japonica, and Brugia malayi.