COMPREHENSIVE ANALYSIS OF GENOMIC STORM (TRANSCRIPTOMIC) DATA, INTEGRATING CLINICAL DATA AND UTILIZING NEW AND OLD APPROACHES USING R AND BIOCONDUCTOR PACKAGES

Abstract

To determine trauma-specific transcriptomic signatures for septic sub-cohorts. In retrospective large-scale data analysis, old and new methods were applied, including lagged correlation between transcripts and clinical subtype counts by integrating over 800 samples from trauma patients. Focusing on novel pathways and correlation methods that were revealed (persistently down-regulated) ribosomal genes and changed time profiles of metabolic enzyme precursors /transcripts. Candidates associated to insulin signaling, including HK3, hinted towards “metabolic syndrome”. Correlation analysis yielded robust results for LCN2 and LTF (r>0.9), but only moderate associations to subtype counts (e.g. top-performing r (Eosinophil, IL5RA)>0.6). Gene Centered Normalization Reduces Ambiguity and Improves Interpretation.