Development of a decision tree to predict correct poses of PD-1/antibody complexes obtained by docking

Introduction: Cancer is a serious health problem. Research has shown that only in 2020 it caused about 10 million deaths and 19.3 million cases. Cancer cells develop due to mutations in proto-oncogene and/or tumour suppressor genes. It is worth mentioning that one of the characteristics of cancer cells is to prevent their elimination by escaping the immune system through the activation of negative regulatory pathways, also known as immune checkpoints. In this work, the PD-1/PD-L1 pathway is highlighted. Currently, one of the most effective treatments is based on the use of monoclonal antibodies (mAbs) as inhibitors of immunological checkpoints. The application of bioinformatics tools in the pharmaceutical industry has enabled the process of research and development of new drugs to be faster, more effective, and less costly. Some of these bioinformatics approaches include molecular modelling, molecular docking, and molecular dynamics simulation. However, accurate identification of correct poses is still an issue.