Fig. 5.

Distribution of SCL predictions from simulated metagenomic reads by PSORTm 1.0.2 and MAGs by PSORTb 3.0.2. PSORTb-MAG refers to PSORTb localization results from protein sequences predicted in the simulated MAGs. PSORTb-Reference refers to PSORTb localization results from protein sequences predicted from the 30 RefSeq bacterial genomes used in the simulated metagenomic dataset. PSORTm-Reads refer to PSORTm prediction from protein sequences directly deduced from the simulated metagenomic reads. Comparison of run time in seconds (top) and number of taxonomic assignment (bottom) as functions of the fraction of reads from the randomly selected watershed discovery sample. Number of taxonomic assignments corresponds to the number of reads successfully analyzed by PSORTm