Bayesian network analysis of immune signaling networks FACS data

Cancer immunotherapy, specifically immune checkpoint blockade therapy, has
been found to be effective in the treatment of metastatic cancers.
However, many patients do not show marked clinical response. Consequently,
elucidating immune system-related pre-treatment biomarkers that are
predictive with respect to sustained clinical response is a major research
priority. Another research priority is evaluating changes in immune
signaling networks before and after treatment in responders and
non-responders. High-dimensional flow cytometry data (FACS,
Fluorescence-activated cell sorting) characterizing immune signaling
network markers in gastrointestinal (GI) cancer patients was used by us to
perform such analyses. We developed a novel computational pipeline to
perform secondary analyses of FACS data using systems biology / machine
learning / information-theoretic techniques and concepts, namely Bayesian
networks and maximum entropy. Application of the pipeline resulted in
elucidation of immune markers, combinations and interactions thereof, and
corresponding immune cell population types, that are associated with
clinical response in the GI cancer cohorts. Future studies are planned to
generalize our analytical approach to different cancer types and
corresponding multimodal high-dimensional datasets.