A   Mean n-mer coverage Pd n-mer ZCAL Context ZYp Y.pestis CO90 64.83 0.99992 0.12 -1.96 Y.pestis KIM 64.23 0.99966 2.87 -1.93 Y.pestis Nepal 64.95 0.99939 5.53 -1.91 Y.pestis Angola 65.28 0.99937 5.78 -1.91 Y.pestisPestoides 64.11 0.99888 10.91 -1.85

B   Contextual   Pd n-mer ZCAL ZCAL ZCAL ZCAL B.globigiiDugway 0.9843 0.02 -0.78 -1.96 -2.76 B.subtilus 0.6810 68.07 615.42 30.14 2.17 Y.pestis CO90 0.9790 1.22 9.99 -1.40 -2.67 Y.pestis KIM 0.9780 1.45 12.02 -1.29 -2.66 Y.pestisPestoides 0.9725 2.7 23.20 -0.71 -2.57 Y.pestis Nepal 0.9760 1.90 16.08 -1.08 -2.62 Y.pestis Angola 0.9755 2.02 17.01 -1.03 -2.62 Y.pseudotuberculosis 0.9080 17.30 154.24 6.12 -1.52 Y.enterocolitica 0.9080 17.30 154.24 6.12 -1.52 Mean signature n-mer coverage = 7.5

Table 2: A: Signature matching results for a pure Y. pestis C090 run showing reference genome database Z values, probability of detecting signature n-mers (Pd) Z values < ˜2.0 are considered consistent with the reference organism being present in the sample within the limits imposed by base calling noise and population structure (“biological noise”). The contextual Z values represent the deviations of the observed Pd from that expected based on the calibration at the same coverage in units of standard deviation of Pd among all other known Yp reference sequences. If Y. pestis C090 was not in the database, KIM strain would be the best scoring, but with Z value > 2.0. However, here the contextual Z value would tell us that the distance from Yp KIM is still well within the distances of Yp among themselves, so treating the sample as an unknown we could state it is very likely a Y. pestis. B) Z value and contextual Z values for Yp and Bg and much lower coverage of the signature n-mers. Here the data for Yp are consistent with 4 known Yp strains. Note the high Z value for B. subtilis, the nearest known neighbor of Bg. It has an ˜80% signature homology with Bg, showing the high sensitivity of this method to small genomic change.