ChEMBL-NTD actives

06 Jun
Published by pylioja

I thought I would just quickly summarise some of the hits that come up in ChEMBL-NTD for the core structures of the compounds we're working with at the moment. The most active compounds, the "near neighbours", contain both a 2-iminothiazolidinone and 2,5-dialkylpyrrole moiety. Chris Southan and Iain Wallace have been helping us track down instances of these compounds in various databases and we're currently compiling a list of desirable compounds to source by any means necessary. More updates coming on this front in the near future.

General Near-Neighbour structure

A serach of the TCAMS dataset, hosted on ChEMBL-NTD, for the 2-iminothiazolidinone core reveals 24 hits while the 2,5-dialkylpyrrole core gives 55 hits. There is some overlap in these searches; the 5 "near-neighbour" compounds originially identified by Paul Willis. An interesting point in the the 2-iminothiazolidinone hits is compound 536110/TCMDC-124424. We tested an analogue of this to probe if our high activities are being caused by the iminothiazolidione component of our molecule. We were pleased to find that OSM-S-55 was relatively inactive against 3D7 (IC50 >5 uM but 100% at 4-8 uM). ChEMBL 536110 is active around 0.5 uM and differs by the presence of 2 methyls on the aryl ring.

OSM-S-55 and ChEMBL 536110

Considering the on-going discussion of rhodanines and their derivatives happening over at Derek Lowe's Pipeline, it might be good to mention them here. We've discussed our compounds about their potential to be PAINS when Jonathan Baell pitched in and again when we've been thinking about our bio results. As mentioned by Zoe, the compounds in the TCAMS dataset were filtered by a "Inhibition Frequency Index" that removed compounds that hit in between 6-25% to act as a safe guard against PAINS. The ChEMBL-NTD does contain 58 rhodanines but only one in the TCAMS dataset (pXC50 3D7 943 nM). Presumably because they were all filtered out by GSK already due to these issues.