Number of sequences and runtime per protein dataset.

All datasets were downloaded from Ensembl Genomes. According to the specifications at Ensembl, datasets specified by “all” represent the super-set of all translations resulting from Ensembl known or novel gene predictions, while datasets specified by “ab initio” include translations resulting from ab initio gene prediction software tools. Such ab initio predictions are based solely on the genomic sequence and not any other experimental evidence, and, therefore, not all predictions represent biologically real proteins. The sequences listed as Seqs valid represent the part of all sequences which meet the necessary conditions for processing (e.g. minimum length of sequence).