Improving the precision and accuracy of predicting radiation exposure

Presentation on radiation biodosimetry by gene expression at Canadian Space Agency June 2, 2021

Mitigating effects of blood disease pathologies that compromise specificity of gene expression signatures for radiation exposure Eliseos J. Mucaki1, Ben C. Shirley1, and Peter K. Rogan1,2 1CytoGnomix Inc., 2 Departments of Biochemistry and Oncology, University of Western Ontario Background. Combinations of expressed genes can discriminate radiation-exposed from normal control blood samples by machine learning based signatures (with 8 to 20% misclassification rates; PMID 29904591). These signatures can quantify therapeutically-relevant as well as acute accidental radiation exposures. The prodromal symptoms of Acute Radiation Syndrome overlap those present in some viral infections. We recently showed that these radiation signatures produced unexpected false positive misclassification of influenza and dengue infected samples (PMID 33299552). Methods. This study investigated recall by previous and novel radiation signatures independently derived from multiple GEO datasets [GSE6874, GSE10640, GSE85570, GSE102971] on common and rare non-malignant blood disorders and blood-borne infections (thromboembolism [GSE19151], S. aureus infection [GSE30119], malaria [GSE117613], sickle cell disease [GSE35007], polycythemia vera [GSE47018], and aplastic anemia [GSE16334]). Normalized expression levels of signature genes is input to machine learning-based classifiers to predict radiation exposure in other hematological conditions. Results and Discussion. Except for aplastic anemia, these confounders modify the normal baseline expression values, leading to false-positive misclassification of radiation exposures in 8 to 54% of individuals. Shared changes predominantly in DNA damage response and apoptosis-related gene transcripts in radiation and confounding hematological conditions compromise the utility of these signatures for radiation assessment. These confounding conditions are known to induce neutrophil extracellular traps, initiated by chromatin decondensation, fragmentation and often, programmed cell death. Ribovirus infections are proposed to deplete RNA binding proteins, inducing R-loops in chromatin which collide with replication forks resulting in DNA damage and apoptosis. To mitigate the effects of confounders, we evaluated predicted radiation positive samples with novel gene expression signatures derived from radiation-responsive transcripts of secreted blood plasma proteins whose expression levels are unperturbed in these confounding conditions. Conclusions. This approach identifies and eliminates misclassified samples with underlying hematological or infectious conditions, leaving only samples with true radiation exposures. Diagnostic accuracy is significantly improved by selecting genes that maximize both sensitivity and specificity in the appropriate tissue using combinations of the best signatures for each of these classes of signatures.