Skip to main content
. 2019 Aug 15;9:11876. doi: 10.1038/s41598-019-48410-y

Figure 5.

Figure 5

Exploring best 4-order combinations and 2-order drug cycling options against drug-resistant M. tuberculosis. In each panel, all pairwise static λ and lytic λ scores among four drugs are given as heatmaps. (top) Heatmaps correspond to best 4-order combinations against drug sensitive or resistant M. tuberculosis. The scores above heatmaps indicate the estimated 4-order interaction scores. (a) eta + fus + pre + van is the lowest ranked 4-order combination in our framework. (b) Since rif and van are similar drugs, we cannot use van against a rif-resistant Mtb strain. Therefore, the lowest ranked 4-order combination shown in (a) is not applicable. When rif and van are excluded, the lowest ranking 4-order combination is clz + fus + inh + sq1. (c) Similarly, since inh and eta are similar drugs, we cannot use eta against an inh-resistant Mtb strain. The best 4-order combination against inh-resistant strain is found as fus + pre + sq1 + van. (d,e) The best 4-order combinations against Mtb strains resistant to rif-inh or rif-inh-mox. (f) The best 4-order combination that includes current Mtb treatment. (bottom) Magenta and green frames show the 2-order combinations nominated for cycling. The two scores correspond to measured static λ and lytic λ scores for each pairwise combination. (g) Best 2-order combination cycling is pre + van and bdq + clz. (hk) Best 2-order cycling options for Mtb strains resistant to indicated drugs. (l) When considered as two 2-order combinations, EIR + clz is split into clz + inh and eth + rif. For each resistance scenario and EIR + clz, the static λ and lytic λ scores in cycling is superior to 4-order combinations.