Welcome to the second AI lounge 探花精选! This time we are happy to have our first presentation from a user given by Rahmad Akbar from the Department of Immunology 探花精选. Please, see the abstract for the talk below and also note the new meeting place.
There will be time for discussion. Looking forward to seeing you at the AI lounge.
Title: Machine learning for predicting antibody recognition in silico
Authors: Rahmad Akbar*, Igor Snapkov*, Milena Pavlovic*,#, Geir Kjetil Sandve#, Victor Greiff*
*, #
Abstract: Antibodies are designed to recognize complex pathogen landscapes such as life-threatening bacteria and viruses and neutralize them via specific binding to these foreign particles. They underlie the principle of vaccination as well as are used as therapeutics in cancer and autoimmunity treatment. Although antibodies are of incredible importance to medical care and public health, the prediction of antibody recognition to a given pathogen remains one of the largest challenges in immunology. Solving this problem will enable the 鈥媔n silico design of antibody based therapeutics. Here, we explore a set of machine and deep learning algorithms to discover and recover motifs/patterns that determine antibody-pathogen binding on simulated sequence datasets. Correspondingly, we describe the challenges that may be encountered in establishing/running machine learning workflows on a high performance clusters (HPC) such as large-scale datasets, parallelization, and efficient resource deployment.
Meeting place: Georg Sverdrups hus, University Library, room 2531 (map)
Meeting time: December 18th, 2018; 15:15-16:15