Drug discovery depends on the discovery of bioactive compounds, in particular small molecules. Virtual or in silico screening is the process of algorithmically screening electronic compounds against mathematical models extracted from a training set. A collaborative project has now been defined to apply a series of models against commercially available compounds to identify potential lead compounds for further examination. Previously screened compounds enable a structure-activity model to be extracted and applied to electronically screen available collections.

The North Carolina State University Exploratory Center for Cheminformatics Research group, funded as part of the NIH Road Map, has developed a web service, ChemModLab, that computes quantitative structure-activity relationship (QSAR) models. The ChemSpider team has delivered our web-based portal, indexing over 16 million chemical compounds from various sources including the scientific literature, chemistry database providers and commercial chemical vendors and screening library providers. This data collection provides an ideal basis for combing the QSAR capabilities of ChemModLab with the compound collection and indexing of ChemSpider to allow identification of virtual hit compounds via in silico screening. A project has been defined to combine the work of the two groups to form a cohesive team to deliver an integrated and optimized system for the delivery of modeling and identification of commercially available compounds for biological testing.

The data sources in ChemSpider linking to screening library vendors and other commercially available compounds presently totals about 7 million compounds but more data will be indexed in the near future. If you are a library vendor please contact us to have your materials indexed as a part of this project in order to expose them to the users of the system. If anyone is interested in potentially participating in the project please contact us directly.

While we are in the early stages of the project at present things are moving quite quickly. Wish us luck!

Stumble it!

2 Responses to “Meshing ChemModLab and ChemSpider to Provide an Optimized Workflow for QSAR Modeling and Identification of Bioactive Compounds”

  1. Jean-Claude Bradley says:

    This sounds like it has a lot of potential. Will our CombiUgi compounds automatically be submitted because they are in ChemSpider and is there an alerting service for “hits”?

  2. Antony Williams says:

    At present we are gathering a proof of concept 1 million structures from different chemical vendors on ChemSpider. We can include a few thousand from Combiugi if you’d like.

    There is no alerting service as yet. It is very early in the work.

Leave a Reply