Skip to main content
. 2016 Aug 25;2(8):e000075. doi: 10.1099/mgen.0.000075

Fig. 2.

Fig. 2.

Magnitudes of errors in proportion estimates of BIB, Pathoscope and naive estimation among uniquely mapping reads (Unique) in strains really present in the experiment (true positives; left) and those not present in the experiment (true negatives; right). The “Unique” method is implemented by simply computing the frequencies of different strain clusters among unique alignments. Lower values indicate better results.