Mesnard, who is a nephrologist, will be working with the Sorbonne Centre for Artificial Intelligence (SCAI), to realise the project that could impact many lives.
“Currently, doctors compare only a tiny part of the donor’s and recipient’s DNA (called Human Leukocyte Antigen) before the transplant to predict if they match.
Improving transplant matches
Post-transplant RNA tests are also conducted to determine if the graft will encounter any problems and also to develop a tailored treatment. However, the current procedures aren’t that efficient, as around 50 percent of kidney transplants fail in 10 years,” he told RFI English.
However, in order to improve the graft efficiency, Mesnard and his team have expanded the pre-transplant matching to other parts of the genome.
They developed a new computational scoring technique called the allogenomics mismatch score, which indicates the number of incompatible amino-acids between a donor’s genome and a recipient’s genome.
“It’s not possible to use this DNA-based method for kidney transplants without the power of AI. One genome consists of 3 billion base pairs. Currently in France, you need five days to analyse a genome. Our goal is to shorten this time to 20 hours by using artificial intelligence,” he said.
It is here that the newly launched AI centre at Sorbonne University will play a crucial part.
Combined research efforts
“Our objective is to bring together researchers from different disciplines under one umbrella and provide a boost to their projects. We want to bring about an alliance between academic knowledge and computational power. SCAI is the AI center where quality research will be done through collaboration,” Professor Gerard Biau, director of SCAI, told RFI.
For this project, SCAI will bring together Mesnard and his team as well as the expertise of the French company Atos - which specialises in high-performance computing - under a single roof.
“SCAI will play the role of a catalyser to accelerate not just this project, but also other initiatives in the fields of science, health, engineering, climate & environment and digital humanities,” Biau said.
Mesnard believes the use of AI in genome sequencing could help raise the success rate of kidney grafts after 10 years from 50 percent to 80 percent.
“By increasing the lifespan of the kidney graft, it would reduce the overall cost for the society as well as improving the quality and quantity of life for the patients,” he said.