to comprehensive catalogues from several other organisms. In contrast, several malarial ePKs were classified at the time of their initial identification several years ago, when the sequences could be compared only to non-comprehensive sets. As an example, 
both P. falciparum MAPKs were identified before the mammalian ERK8/7 enzymes were discovered, and the closest sequences available at the time were those of the ERK1/2 family. Secondly, it has been reported that BLAST performs poorly in assigning orthology between human and C. elegans genes. This is because of extensive independent gene duplication on the lineages leading to the two organisms. Humans and P. falciparum are much more distantly related and there has been extensive gene duplication on the human side. Our data support the view that reliable assignments of orthology between genes in distantly related species might only be assigned through the construction of phylogenetic trees and suggest that comparisons based on BLAST must be interpreted cautiously. FIKK, a novel, Apicomplexa-specific group of ePK-related proteins That only 65 typical ePKs were identified in this search is somewhat surprising, as Saccharomyces cerevisiae, whose genome is half the size of the P. falciparum genome, encodes approximately twice as many enzymes of this family. cule in the catalytic cleft. However, one, and sometimes two glycine residues are present in subdomain I of the FIKK sequences. This is also the case in a number of enzymes with demonstrated protein kinase activity from many organisms , and it is clearly established that ATP binding and phosphotransfer ability is not dependent on the presence of a Glycine triad. Although lacking the Glycine triad, all FIKK sequences possess an N-terminal extension, with a conserved tryptophan residue in the Acacetin web region that corresponds to subdomain I. One of the FIKK sequences is represented in PlasmoDB as two contiguous ORFs separated by a gap. This is presumably due to erroneous prediction: alignment with other FIKKs clearly shows these sequences represent two parts of a single member of the FIKK family rather than two separate genes. Furthermore, RT-PCR across the two predicted ORFs demonstrates that both sections are present on the same mRNA molecule. Interestingly, sequencing of the RT-PCR product showed that the open reading frame in the cDNA is interrupted by an in-frame stop codon, which is presumably the cause of the misprediction of the gene structure. That this sequence is cDNA than genomic is ascertained by the presence of an intron near the 3’end. Whether PF14_0733/4 is a transcribed pseudogene, or whether a protein can be produced by readthrough of the internal stop codon as has been documented for another P. falciparum gene, remains to be determined. In any case, it appears there are only 20 FIKK sequences in the genome, instead of the 21 that were counted originally. In addition to the residues conserved in typical ePKs, several amino-acid motifs are fully conserved in all members of the FIKK family. These can be used to define signature motifs, which allowed us to perform a number of motif searches in various databases, to determine whether members of this ePK-like family are present in other organisms. Interestingly, sequences containing such motifs could be retrieved only from Apicomplexan species: 20 sequences in the P. falciparum genome, one in P. berghei, one in P. yoelii, one PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19793655 in P. knowlesi and one in P. vivax. In contrast, no FIKK family
