Fig. 2

Reciprocal best hits found by diamond and mmseqs. The figure shows results using the E. coli proteome as reference. Results with the other reference proteomes showed similar tendencies (see supplementary document). The proportion of RBH found is comparable to those found by blastp when the proteomes involved are very similar (high Genomic Similarity Scores, GSS). This proportion reduces with the GSS. As expected, low sensitivities reduced the proportion of RBH found. Note that the improvement between diamond’s very-sensitive and ultra-sensitive settings is modest compared to their relative runtimes (Fig. 1)