Meeting with LIFEWARE
The LIFEWARE project-team joined the Inria Saclay - Île-de-France research centre in January 2016. Between hardware and software, LIFEWARE seeks to shed light on life software by trying to understand its methods using engineering methods. Meeting with François Fages, Grégory Batt and Sylvain Soliman.
Understanding the logic of life: a vast programme
The LIFEWARE team, which was created in January 2014 as a follow-up to the CONTRAINTES project and became a project team in April 2015, is made up of computer scientists and logicians who use logical formalisms to model interactions and behaviour.
LIFEWARE seeks to understand the different cell processes as computer programs - seeing cells as machines, all biochemical reactions as programs and modelling these processes - with applications for cancer, for example, which is one of the consequences of a malfunction of cell proliferation decisions.
"In a way, we are looking to see how the cell computes", François Fages explains.
Going further and questioning evolution
When the model does not work for a certain type of cells, its structure then needs to be reviewed as do known supposed molecular interactions, and new predictions on these interactions need to be carried which are then verified through experimentation in biology.
In the past, the team members developed techniques to describe the systems and their behaviour, and to calibrate the models on experimental data.
"But we want to go further, we are not merely looking to describe", François Fages tells us.
What is the set of functions that can be carried out by biological circuits? Where do these circuits come from? What is missing to get them to do something else? LIFEWARE's real challenge at present is to try to understand, over the long-term, why they are made and why nature is where it is.
The collaboration between biology and computer science
François Fages is endeavouring to get computer science and biology to work together : "I have always had a collaboration policy with biologists that is quite opportunistic and focuses on concrete problems. We don't have the same languages - it is not a question of trying to make them match, but of working together on a concrete problem".
The LIFEWARE team has notably worked with oncologists on cancer chronotherapy and with biologists on GPCR signalling in order to produce models thanks to the Biocham software - modelling software that has been developed for 13 years. In particular, the collaboration between the LIFEWARE team and Eric Reiter (Inra) earned them a paper with Robert Lefkowitz, who received a Nobel Prize for Chemistry in 2012.
"Through computer modelling we have really learned from the biology model, at the highest level," François Fages underlines.
One of the team members, Grégory Batt, has also worked with Pascal Hersen, physicist at the Université Paris 7. Together, they succeeded in controlling the expression of a gene by changing the osmotic pressure in a microfluidic circuit. Grégory is currently working on cell-to-cell variability, which consists in not simply modelling the behaviour of a process within a cell but also the variability of this process. "It is very important work because, in all of our experiments, we can see that not all cells react in the same way", François Fages stresses. Grégory Batt is working with the Institut Pasteur on the creation of a group named InBio around these themes, as an Inria exploratory action, whilst also remaining with the LIFEWARE project.
A methodology that applies to biology - but not only biology
The members of the team are from the logic optimisation school and apply these methods to biology, but continue to work on constraint and optimisation problems.
Sylvain Soliman: "Even if our name has changed, a small part of the CONTRAINTES project still lives on in LIFEWARE. From time to time we continue to work on optimisation problems that have nothing to do with biology (placement optimisation, energy optimisation for underground networks...). The link to biology is to be found in the methodology; we use the same algorithms but with a different application."
François Fages explains: "Broadly speaking, we come from logic, constraints and the solving of combinatorial problems, which are discrete problems. By taking an interest in biology, we have imported discrete methods into that area. However, very quickly, we needed to work on continuous models that, in return, we imported into the field of logic. The two thrive on each other. "
What does "signalling" mean?
Signalling is the transfer of information between something that happens outside the cell, that is captured by a transmembrane receptor, and an action that will generally be decided further in the cell's nucleus. Between this outside signal and the decision as to which action will result from it, there is a whole series of reactions - this complex process is what we call signalling.