Fig. 2

Evaluation of the SuRankCo rankings on the E. coli test data. a shows the distribution of SuRankCo contig scores. They form two clusters based on the high correlation of target scores in this data set. Clusters are skewed due to classification probabilities incorporated into the SuRankCo contig scores. b shows a scatterplot comparison of the ranking and the pIdent of Blast matches against the E. coli taxon. High and low density areas are indicated in red and blue, respectively. Data points below 95 % pIdent are not shown to improve the scaling (25 of 11336)