Systems and methods for MHC class II epitope prediction

A method of identifying one or more immunodominant epitopes, comprising:

• a) generating, using one or more processors, a set of candidate antigens from one or more input genome sequences, wherein the one or more input genome sequences is derived from a target pathogen, a commensal microorganism, or a diseased cell;
• b) generating, using the one or more processors, a ranked set of antigenic epitopes from the candidate antigens using a deep neural network, wherein the deep neural network is trained using a set of MHCII-presented peptides bound to antigen presenting cells, wherein the antigen presenting cells are dendritic cells;
• c) detecting whether one or more of the antigenic epitopes is present in a sample from a subject suffering from an infection, autoimmune disease, allergy, or cancer;
• d) formulating an immunological composition comprising one or more of the identified epitopes;
• and
• e) identifying one or more immunodominant epitopes by:
• i) administering the immunological composition to a first population of immune cells to allow expression of the one or more identified epitopes, and wherein the first population of immune cells comprise a detectable reporter gene;
• ii) expressing in a second population of CD4+ immune cells a TCRaβ library expressing TCRaβ pairs identified in a subject suffering from an infectious disease, autoimmunity, cancer or allergy;
• ii) incubating the first population of immune cells with the second population of immune cells;
• iv) sorting immune cells positive for the detectable reporter gene;
• and
• v) identifying peptides bound to MHCII in immune cells positive for the detectable reporter gene, wherein the peptides identified comprise the immunodominant epitopes.