apply ML algorithms to target the engaged T cells in T1D and finds markers in the TCR repertoire level

Type 1 Diabetes (T1D) is an autoimmune disease resulting from T cell destruction of insulin-producing β-cells within pancreatic islets.
Genetics plays a major role in the etiology of T1D.
most of them related to the HR which encode proteins presenting antigen to CD4+ T cells.
in the theory of positive and negative selection, Only T cells that express a TCR that binds a peptide presented by MHC (pMHC) with sufficient affinity will survive positive selection.
then, in negative selection, TCR that interacts with too high affinity will be eliminated.
in the T1D, we can infer that those clones missed in the negative selection are the cause of the autoimmune state.
our goal is first to support the results that show the effect of the DR3 and DR4 on the T cell repertoire gene expression. after that, we need to collect additional samples to train an ML model to predict whether one`s has T1D or not.
in case we will find such a model, we will try to trace his decision-making and get a better understanding of the "clones" that are involved in this autoimmune state.