Analyses of over 700 transcripts with machine learning yielded about 77% classification accuracy, highlighting immune dysregulation through cell-specific signals.