Transcript Using TB Mobile to Predict Potential Targets for TB hits

Using TB Mobile to Predict Potential Targets for TB hits from Phenotypic Screening
Sean Ekins1,2 and Alex M. Clark3
1 Collaborative
Drug Discovery, 1633 Bayshore Hwy, Suite 342, Burlingame, CA 94010, 2 Collaborations in Chemistry, 5616 Hilltop Needmore Road, Fuquay Varina, NC 27526, U.S.A., 2 Molecular Materials Informatics, 1900
St. Jacques #302, Montreal, Quebec, Canada H3J 2S1., (S. Ekins can be contacted at [email protected])
Introduction
In order for us to learn from the extensive prior literature we have
collated information on molecules screened versus Mycobacterium
tuberculosis (Mtb) and their predominantly experimentally confirmed
targets which has been made available in the Collaborative Drug
Discovery (CDD) database. This dataset contains data on target,
essentiality, links to PubMed, TBDB, TBCyc and human homolog
information.
component was used to rank the content in TB Mobile of molecules with
known targets. We have used this as an example of inferring potential
targets. The molecules are ranked and labelled with their target/s. It
should be noted that such data is far from definitive as these published
compounds have not been tested versus all Mtb targets and it is possible
the same compound may be active against more than one target.
The development of mobile cheminformatics apps could potentially
lower the barrier to drug discovery and promote collaboration.
Therefore we have used this set of over 700 molecules screened
versus Mtb and their targets to create a free mobile app (TB Mobile)
that displays molecule structure and links to the bioinformatics data.
By input of the molecule structure the user can perform a similarity
search within the app and infer potential targets. In addition one can
search by targets to retrieve compounds known to be active.
Results
Figure 2 and Table 1 suggest the 11 hits from GSK may be targeting a
limited array of targets. This could be due to limitation of the underlying
data in TB Mobile biased towards those with larger numbers of
molecules. Some molecules have features one sees repeatedly e.g.
GSK353069A looks like a dhfr inhibitor. We have not performed any
experimental verification of these predictions and present our
observations openly to demonstrate the potential utility of TB Mobile and
foster collaboration. Compound availability is however unclear.
Methods
The development of the TB mobile is described elsewhere (Ekins et
al.,submitted), we now focus on evaluation of a recent dataset of 11
compounds that came out of whole cell screening by GSK (Ballell et
al 2013, in press) to demonstrate how the app can be used.
The workflow used is as follows: First the 11 molecules were drawn in
the MMDSapp and exported into the TB Mobile app (an example of
app-to-app communication, Figure 1). The similarity searching
References
L. Ballell, R. H. Bates, R. J. Young, D. Alvarez-Gomez, E. Alvarez-Ruiz,
V. Barroso, D. Blanco, B. Crespo, J. Escribano, R. Gonzalez, S. Lozano,
S. Huss, A. Santos-Villarejo, J. J. Martin-Plaza, A. Mendoza, M. J.
Rebollo-Lopez, M. Remuinan-Blanco, J. L. Lavandera, E. Perez-Herran,
F. J. Gamo-Benito, J. F. Garcia-Bustos, D. Barros, J. P. Castro, N.
Cammack, Fueling Open-Source drug discovery: 177 small-molecule
leads against tuberculosis ChemMedChem 2013.
Figure 1. Workflow from sketching molecules in MMDS mobile app to exporting and opening with TB Mobile
Table 1. Predictions for GSK compounds from phenotypic Mtb screening
Molecule
Name
H37Rv MIC uM
Predicted targets
GSK153890A
0.47
inhA, dxs1
GSK163574A
0.76
dxs1, inhA
GSK1589673A
0.25
inhA
GSK1829820A
0.19
sahH, nedR
GSK1829731A
0.19
dxs1
GSK2200150A
0.38
inhA
GSK353069A
0.13
sahh (looks like dhfr)
GSK358607A
0.7
inhA
GSK749336A
0.25
inhA
GSK888636A
0.94
dxs1
GW623128
0.47
glf, def
Data sources
TB Mobile for iOS
https://itunes.apple.com/us/app/tbmobile/id567461644?mt=8
TB Mobile for Android
http://play.google.com/store/apps/details
?id=com.mmi.android.tbmobile
CDD
www.collaborativedrug.com
Figure 2. screenshots of similarity searching in TB mobile for each of the GSK molecules in the Ballel et al 2013 paper.