Extended Methods and Biological Networks using Explainable AI algorithm Iterative Random Forest

The rapid pace of increasing biological data may require a combination of high performance computing, machine learning, data analytics, graph theory, and systems biology-based processes to best further scientific research. Almost all biological functions operate in conjunction with the system they are in and to best understand them, systems biology studies the function in the context of its system. To handle the highly interacting complexities within these systems, new high performance and human understandable algorithms should be utilized. To leverage the explainable AI capabilities of machine learning model iterative Random Forest (iRF), a high performance computing capable implementation is introduced. This new implementation enables iRF-LOOP and iRF Cross-Layer, new network methods for omics interaction analysis for within- and across-omic relationships. By using these methods together, iRF-based heterogeneous networks with omics layers provide a systems biology view. Using these methods with public data for model organism Arabidopsis thaliana, a large-scale network resource is produced for hypothesis generation and research utility.