Our scientists and many biomedical researchers across the world use small molecules called chemical probes to understand the complex interplay between proteins in cells.

These molecules are designed to bind to a target in a cell, most commonly a protein, and alter or inhibit its function. They can be used as investigational tools or prototype drugs, and ultimately help to discover new treatments for a range of diseases, including cancer.

But the use of poor-quality compounds as chemical probes, often based on out-of-date information, is widespread among the global biomedical research community. It has frequently compromised the robustness of research findings and led to incorrect conclusions.

Scientists at The Institute of Cancer Research have for some time been working to help researchers choose their chemical probes more wisely.

In 2015, we were co-founders of the Chemical Probes Portal (described in a Nature Chemical Biology article), an open-access online resource which provides expert reviews of chemical probes for use in biomedical research and drug discovery.

To be listed on the Chemical Probes Portal, a chemical probe needs to be peer-reviewed by a member of the Scientific Advisory Board, made up of experts from the medicinal chemistry, pharmacology and chemical biology communities.

Users of the portal can now sort through close to 200 high-quality chemical probes, and select the most suitable for their needs, quickly and efficiently.

Although the expert-review approach is valuable and the portal is widely used, a limitation is its far-from-complete coverage of the full range of proteins in human cells.

More recently, our researchers have also launched a highly innovative web resource called Probe Minerthat provides a very complementary approach to selecting chemical probes.

Jointly conceived by Dr Albert Antolin, Professor Bissan Al-Lazikani, Professor Ian Collinsand Professor Paul Workmanfrom the ICR, Probe Miner uses data science and statistical analysis to sort through hundreds of thousands of actual or potential chemical probes.

It ranks their suitability for use against a large number of human proteins based on quantitative statistical assessment according to critical criteria or fitness factors that were defined previously by ICR scientists.

Probe Miner uses data about probes and proteins from canSAR, a knowledgebase also created by researchers at the ICR, which brings together and stores data across biology, chemistry, pharmacology, structural biology, cellular networks, clinical annotations and more.

Looking at the interactions between the chemical probes and the target protein of interest, Probe Miner scores each probe for six fitness factors including their potency, selectivity, cell activity and other features and ranks their suitability for use to study target proteins.

This is quite an undertaking: within canSAR, there are 355,305 potential chemical probes and 2,220 possible human protein targets. But Probe Miners computational methodology can quickly and efficiently produce a quantitative, data-driven assessment of each chemical probe against a given target, and accurately suggest the best probes for use in the laboratory.

Since its conception in 2017, Probe Miner has had more than 4,000 users including researchers from biotech companies such as BenevolentAI, pharmaceutical companies such as GSK and Novartis, and academic institutions such as the Broad Institute of Harvard and MIT, and the University of Oxford.

At the ICR, we are leading in innovative cancer informatics. From mapping the paths of cancer evolution, to the design and discovery of novel drugs, to the precision tailoring of radiation therapy. Informatics research at the ICR spans bioinformatics and computational biology, biostatistics, mathematical biology, in-silico medicinal chemistry, digital pathology and computational physics.

Read more

The creators of Probe Miner are now looking at how to advance and further improve the scope and usability of the site. As well as presentational changes, the team is looking to increase the amount of data accessible to Probe Miner.

ICR scientists have recently called for more integration between Probe Miner and other resources developed to help researchers use chemical probes more wisely, including the Chemical Probes Portal. The two approaches are highly complementary.

With better integration between online resources, the potential for more effective probe selection will be greater than ever before.

As Dr Albert Antolin, Sir Henry Wellcome Postdoctoral Fellow in the Division of Cancer Therapeuticsand the Department of Data Science, explains:

We are bringing together the data and methodology that facilitates objective assessment and selection of chemical probes for scientists who are not experts in chemical biology.

This way we will hopefully address some of the problems around data robustness that we are currently facing in biomedical research, and facilitate the translation of biological discoveries into new drugs for patients.

Original post:
Science Talk - Enhancing the selection and use of chemical probes in cancer research using innovative data science - The Institute of Cancer Research