Minding the gaps in tuberculosis research

- We describe the analysis of 70 years of mouse TB in vivo data.
- We visualize a 'valley of death' of approximately 40 years in TB research.
- We provide suggestions of how to prioritize the 1000s of in vitro hits.
- We suggest the need for more in vivo testing in mice.
- The machine learning model approach used with this in vivo data may be applicable to other datasets.