October 15, 2021
Bulletin interne de l'Institut Pasteur
Scientists from the Institut Pasteur and the Massachusetts Institute of Technology (MIT) have developed a technique for the rapid reconstruction of whole genomes, such as the human genome. Inspired by natural language processing, which considers whole words rather than letters, the scientists set out to represent genomic data using blocks of letters. Using this method, they were able to reconstruct a human genome in record time and in a much less computer-resource-intensive way – the reconstruction process is a hundred times faster than existing state-of-the-art assembly techniques and only requires a fifth of the genomic resources.