Background: Despite the continuous production of genome sequence for a number of organisms,/nreliable, comprehensive, and cost effective gene prediction remains problematic. This is particularly/ntrue for genomes for which there is not a large collection of known gene sequences, such as the/nrecently published chicken genome. We used the chicken sequence to test comparative and/nhomology-based gene-finding methods followed by experimental validation as an effective genome/nannotation method./nResults: We performed experimental evaluation by RT-PCR of three different computational gene/nfinders, Ensembl, SGP2 and TWINSCAN, applied to the chicken genome. A Venn diagram was/ncomputed and each component of it was evaluated. The results showed that de novo comparative/nmethods can identify up to about 700 chicken genes with no previous evidence of expression, and/ncan correctly extend about 40% of homology-based predictions at the 5' end./nConclusions: De novo comparative gene prediction followed by experimental verification is/neffective at enhancing the annotation of the newly sequenced genomes provided by standard/nhomology-based methods.

This work was supported by grants from the Jérôme Lejeune,/nChildcare and Désirée and Niels Yde Foundations, the European Union, the/nSwiss National Science Foundation and the NCCR Frontiers in Genetics. E/nEyras is funded by the Institució Catalana de Recerca I Estudis Avançats/n(ICREA). R Guigo's lab acknowledges support from grants from the/nNational Plan for R&D (Spain), QLK3-CT-2002-02062 from the European/nCommunity and HG003150-01 from the National Institutes of Health./nResearch on gene prediction by vertebrate genome comparison in the/nBrent lab is supported by grant R01 HG02278 from the National Institutes/nof Health. This work was also supported by EMBL and the Wellcome Trust.