Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Already on GitHub? Sign in to your account

hERG Pharmacophore Overlay #188

Open
mattodd opened this Issue Apr 24, 2014 · 3 comments

Comments

Projects
None yet
3 participants
Owner

mattodd commented Apr 24, 2014

hERG data were obtained for two OSM Series 4 compounds (MMV669844 and MMV670944) indicating possible issues. In the latest OSM meeting:
http://malaria.ourexperiment.org/osddmalaria_meeting_/9470
it was suggested someone might want to do an overlay of these compounds vs. what is known of (the) hERG pharmacophore(s). @murrayfold volunteered in the meeting. Data on the compounds are here:
http://malaria.ourexperiment.org/biological_data/9562/hERG_Data_for_MMV669844_and_MMV670944.html
Discussion can take place below, with any modelling outputs going to:
http://malaria.ourexperiment.org/pharmacophore_modelling_

@murrayfold murrayfold was assigned by mattodd Apr 24, 2014

Owner

mattodd commented Jun 29, 2014

More data should be coming #211 . This issue/analysis still interesting/neede, however.
Chembl has been analysed for what it tells us about herg @madgpap
http://pubs.acs.org/doi/abs/10.1021/ci400308z
Seems odd to me that there is no pharmacophore model, implying it's hard (e.g. promiscuous target).

Owner

mattodd commented Jul 1, 2014

Chris Swain from Cambridge MedChem Consulting wrote by email (for public consumption) the following very useful remarks:

"I’ve written a page on HERG, and compiled a database of data taken from the literature, it also includes some general ideas on reducing HERG activity based on a matched pair analysis I did.

I did try to build a global predictive model using the 1600 molecules I have collected. To be honest whilst it might be useful for filtering screening collections it lacks the resolution needed to support SAR studies.

I gave the dataset to Bobby Glen in Cambridge and he looked at published models. http://dx.doi.org/10.1002/minf.201000159

The bottom line

"The work presented here shows the considerable variability in results that may be obtained by training and validating the same hERG blocker classifiers on different datasets. This is highlighted by the considerable reduction in model performance observed on our dataset with the models proposed by Dubus et al and by Thai and Ecker"

My own feeling is that you need to build local models to support SAR studies, however the empirical rules described on my web page are often enough to give a good steer."
@murrayfold
@PaulWillisMMV

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment