Abstract
Within the field of forensic pathology, determination of the cause of death depends upon identifying physical changes in the corpse or finding diagnostic laboratory abnormalities. When such perturbations are absent, definitive assignment of a cause of death may be difficult or impossible. An example of such a problem is sudden infant death syndrome (SIDS), a common cause of neonatal mortality that does not produce physical findings or laboratory abnormalities. Although respiratory failure as a cause of SIDS represents the most widely held hypothesis, sudden cardiac death and hyperthermia have also been advanced as possible causes. We hypothesize that each of these physiological stresses would produce a different pattern of premortem gene expression and that these patterns of gene expression would remain evident in tissues collected postmortem. If these patterns were sufficiently distinctive, they could be used to identify the cause of death. Using an infant mouse model, we compared gene expression patterns in liver tissue after sudden death, lethal hyperthermia, and lethal hypoxia. Each of these conditions produced readily distinguishable differences in gene expression patterns. With the K-nearest neighbor classification algorithm, only 10 genes are necessary to correctly classify samples. If the liver tissue was not harvested immediately after death, additional alteration in gene expression patterns resulted; however, these alterations did not affect the group of genes used to classify the samples. Our findings suggest that gene expression analysis from tissues collected postmortem may provide useful clues about certain physiologic stresses that may precede death.