Identification of Attractive Drug Targets in Neglected-Disease Pathogens Using an In Silico Approach (original) (raw)
Figure 3
The sensitivity of target rankings to changes in weighting.
Using the M. tuberculosis genome as an example, we determined the fraction of genes matching an attribute/query in a set of curated targets (validated chemically and/or genetically) and in the entire genome. (A) The results are shown for each attribute used in Query 5 of Figure 2. Values are log(Observed/Expected), where Expected is the fraction of genes in the genome that have the attribute and Observed is the fraction of curated targets that have the attribute. (B) We analyzed the stability of the final ranked list when the weight of a single attribute is changed. As an indication of stability, we plot the percentage of curated targets among the top 200 genes as the weight of each attribute is changed from minus-100 to plus-200.