Interview with Philippe Roche

Published on

The Marseille Cancer Research Center celebrates its 50th anniversary ! -

Interview with Philippe Roche, by the Communication Cell, June 2024


Can you summarize this article in a few words?

The development of bioactive molecules or drugs generally begins with fragment screening to identify primary hits. However, these initial compounds need to be optimized to enhance their affinity and specificity. This optimization phase involves multiple cycles of chemical modifications during which extensions are added to the initial fragment.

ChemoDOTS is a web server that generates all easily synthesizable compounds (in one or two steps) from an initial fragment, integrating commonly used medicinal chemistry rules in the pharmaceutical industry. The server is straightforward to use. Typically, users must i/ import the initial fragment or draw it, then ii/ choose among automatically detected reactive functions to which the extensions will be added. In return, the server provides a list of compatible chemical reactions that can be selected to automatically generate all synthesizable compounds from the initial fragment. Physico-chemical filters can be applied to extract a subset of compounds with specific properties. The principle is summarized in the diagram below.

How does the ChemoDOTS server represent a significant advancement in the field?

This server allows for the rapid generation of a large number of compounds, representing a potential source of optimized molecules that can be easily synthesized. Similar approaches had already been developed by several teams, including ours, using custom scripts, but they were mainly dedicated to chemoinformaticians. This server provides a simplified access to the scientific community and thus opens the way for its use by a broader audience. Moreover, since its launch, the feedback on this server has been very positive.

What challenges were encountered and how were they overcome?

We have been using this approach in the team for several years. A large number of developments were made by Laurent Hoffer, an excellent postdoc who spent six years in our team. But with Xavier, we had the idea of developing a more user-friendly server for several years. A first version was quickly developed by Samuel Granjeaud.

More recently, thanks to the arrival of a new PhD student, Guillaume Charifi, who is particularly skilled at optimizing computer code, we were able to develop a new version that is very user-friendly and fast. Tens of thousands of molecules can be generated in just a few seconds. One of the major difficulties during these years was finding the right student to finalize this project, because developing a web server is not necessarily considered a research project in itself, even though there is strong demand from the scientific community.

What future developments can be anticipated as a result of this research?

We already use this server (or its command line versions) in most of the team’s projects in hit-to-lead optimization programs. The organization of our team, with chemists integrated into research projects, is particularly suited to this approach. It is currently possible to improve an initial hit in just a few weeks. New features are already under development (addition of new reactions, other filters, the possibility of starting from non-functionalized molecules, etc.).

We also apply similar approaches for the design of covalent probes, which have become popular in recent years with the approval of several drugs, particularly in oncology.

Are there simple analogies that could help non-scientists in understanding the significance of this work?

An analogy can be made with a Lego game. Drugs (or any molecule) can be considered as the assembly of several elementary bricks. The ChemoDOTS approach allows for building a large number of new compounds by adding elementary bricks to an initial brick, aiming to obtain more effective molecules, as summarized in the diagram below.


A bit more about you: your background, what led you to science.

My scientific career has not followed a linear path, but it has been enriched by numerous meaningful encounters that have frequently guided my thematic and geographical shifts. I started my university studies in Orleans, then in Toulouse, where I was originally interested in organic chemistry. A decisive encounter with Professor Jean René Pougny led me to pursue a Master’s in microbiology, however I managed to do both of my practical internships in synthetic chemistry. I then had the chance to arrive at the right place at the right time for my thesis on plant-bacteria symbiosis between legumes and Rhizobia. During these years, I had a very strong bond with Patrice Lerouge, a CNRS researcher. This led to the characterization of the first nodulation factors and a very productive period that allowed me to obtain a position at CNRS in “Plant Biology”.

I continued with a postdoctoral fellowship at the “Complex Carbohydrate Research Center” in Athens, GA (USA). Upon my return to France, wanting to focus on the characterization of nodulation factor receptors, I took a new turn in my career by training in molecular modeling and dynamics, notably in the laboratories of David Perahia in Orsay and David Stuart in Oxford. Over the years, I definitively moved away from nodulation factors to focus on modeling protein-protein and protein-ligand interactions.

I then joined Françoise Guerlesquin’s team in Marseille before joining the CRCM in 2012 in Xavier Morelli’s team where I am still today. I take great pleasure in being part of the iSCB team because we share many values with Xavier. I am convinced that he would embrace my motto, which could be that of the iSCB team: “Alone we go faster, but together we go further.”

Tell us more about your hobbies and interests outside of your professional life, particularly your involvement in music.

Although I am not from a family of musicians, I immersed myself in the musical world at a very young age. I studied slide trombone at the “Conservatoire” of Orleans, but I was particularly interested in playing in orchestras. There too, I was fortunate enough to play alongside some very talented musicians in various ensembles. I appreciate the versatility of the trombone, which has allowed me to play in a wide range of musical styles: classical, street music, bandas, Dixieland orchestras, jazz big bands, salsa groups, and rhythm-and-blues groups. Obviously, I traveled with my trombone during my two postdocs. Unfortunately, I had to take a break for over fifteen years due to severe tinnitus.

Recently, I resumed playing with the CNRS Big Band, the University symphony orchestra (OSAMU), and some local bands. Currently, my musical activity is again on hold, but I still feel a void when attending concerts … so who knows.


To know more:

Hoffer L, Charifi-Hoareau G, Barelier S, Betzi S, Miller T, Morelli X, Roche P. ChemoDOTS: a web server to design chemistry-driven focused libraries. Nucleic Acids Res. 2024 Apr 30:gkae326. doi: 10.1093/nar/gkae326. Epub ahead of print. PMID: 38686808.