Human and nematode orthologs - lessons from the analysis of 1800 human genes and the proteome of Caenorhabditis elegans

Recently, we have defined and analyzed over 1800 orthologous human and rode nt genes. Here we extend this work to compare human and Caenorhabditis eleg ans coding sequences. 1880 human proteins were compared with about 20000 pr edicted nematode proteins presumably comprising nearly the complete proteom e of C. elegans, We found that 44% of human/rodent orthologs have convincin g nematode counterparts. On average, the amino acid similarity and identity between aligned human and C. elegans orthologous gene products are 69.3% a nd 49.1% respectively, and the nucleotide identity is 49.8%. Detailed inves tigation of our results suggests that some nematode gene predictions are in correct, leading to erroneous pairing with human genes (e.g. calcineurin an d polymerase II elongation factor III). Furthermore, other proteins (i.e. h omologs of human ribosomal proteins S20 and L41, thymosin) are missing enti rely from the nematode proteome, suggesting that it may not be complete. Th ese results underscore the fact that metazoan gene prediction is a very cha llenging task and that most computer-predicted nematode genes require suppo rting evidence of their existence from comparative genomics and/or laborato ry investigation. (C) 1999 Elsevier Science B.V. All rights reserved.