Identification of "missing" metabolic proteins of Plasmodium falciparum: a bioinformatics approach

The genome of Plasmodium falciparum, a malarial parasite, is atypical as many of the encoded proteins have diverged extensively from their homologues in other organisms. Hence homology-based information transfer is not entirely successful and presently, proper function annotation is unavailable for over 50% of the proteome. It has been hypothesized that enzymes participating in nearly 69 metabolic steps are not yet identified. In this paper we report detection of some of the "missing metabolic enzymes" of P. falciparum. Our approach for remote homology detection to recognize the "missing" P. falciparum enzymes employs multiple profiles for every protein domain family. A blind assessment of the approach to recognize known metabolic proteins of P. falciparum resulted in 94% success rate. Using this approach we have successfully recognized 14 of the "missing" enzymes. Results of protein fold recognition and information from microarray and protein-protein interaction datasets are consistent with our predictions. In a few cases we also provide the list of functionally important residues extrapolated on the basis of homology.