Improving gene models of Chlamydomonas reinhardtii by mass spectrometric data from Proteomics experiments

Michael Specht1, Mario Stanke2 and Michael Hippler1

1) Institute of Plant Biochemistry and Biotechnology, University of Münster, Germany 2) Institute of Microbiology and Genetics, University of Göttingen, Germany

We present the Genomic Peptide Finder (GPF), a tool that matches de novo predicted amino acids sequences to the genomic DNA sequence of an organism while a) allowing for errors within the predicted sequence and b) deducing intron-splits on the DNA level. The powerful combination of de novo prediction and subsequent peptide alignment onto genomic DNA via GPF allowed for deduction of more than 12,000 gene model-unbiased, DNA-aligned peptides from a total of 17 Proteomics experiments which were originally not intended for gene model prediction. About 25% of the alignments contain an intron split, which in some cases appear within a single coding nucleotide triplet. These alignments were passed to the ab initio gene model prediction program AUGUSTUS as external exon/intron boundary hints, aiding in the prediction of the Chlamydomonas reinhardtii v5.0 gene models.

e-mail address of presenting author: mhippler@uni-muenster.de