Rule-based machine learning for prediction of Macaca mulatta SIV-vaccination outcome using transcriptome profiles

Abstract: One of the reasons, why the development of an effective HIV vaccine remains challenging, is the lack of understanding of potential vaccination-induced protection mechanisms. In the present study, Rhesus Macaques (Macaca mulatta) gene expression profiles obtained during vaccination with promising candidate vaccines against Simian Immunodeficiency Virus (SIV) were processed with a rule-based supervised machine learning approach to analyze the effects of vaccine combination treatment. The findings from constructed rule-based classifiers suggest that the immune response against SIV builds up throughout the immunization procedure. The upregulation of three genes (NHEJ1, GBP7, LAMB1), known to contribute to immune system development and functioning, cellular signalling, and DNA reparation, during or after vaccination boost appears to play an important role in the development of protection against SIV. What is more, the data suggest that the mechanisms of protection development might be dependent on the vaccine type providing a plausible explanation for the difference in effect between vaccines. Further studies are necessary to confirm or disprove our preliminary understanding of the vaccination-induced protection mechanisms against SIV and to use this information for rational vaccine design.