Identifying At-Risk Patients
At the “Critical Assessment of Massive Data Analysis” conference held recently in Vienna, John Storey, from the Lewis-Sigler Institute for Integrative Genomics at Princeton University, gave a talk on dissecting the inflammatory complications in critically injured patients by within-patient gene-expression changes.
Trauma is a major killer, as it can lead to sepsis, infection, and multiple organ failure. But it has been difficult to determine which hospitalized patients will succumb to these severe sequelae and which will have a smoother recovery. Dr. Storey participated in a longitudinal study to try to identify which genetic pathways are associated with more direct clinical outcomes as well as exactly when these molecular signatures must be detected to be useful in determining clinical treatments.
Using mRNA from total blood leukocytes, they found that a full one-quarter of the human genome exhibits changes in expression levels during the early stages of post-trauma (40–80 hours post-trauma). Dr. Storey specializes in quantitative genomics; his statistical challenge in this study was to correlate gene expression over multiple time points, in multiple patients, with multiple and different clinical presentations.
“It took me a long time to appreciate just how extensively population heterogeneity and other unmeasured sources of variation introduce unwanted and systematic biases into the statistical analysis of gene-expression studies,” Dr. Storey noted.
“I view this problem as being as or more problematic than the population stratification problem in genome-wide association studies (referring to allelic differences in subpopulations that may be due to ancestral, rather than disease associated, differences).”
His innovation was to correlate the within-patient change in gene expression with the severity of the clinical outcome, rather than using absolute expression values as has traditionally been done. By measuring the change in expression within each patient, regardless of the patient’s baseline values, this method leads more easily to a clinical translation of the results.
Hara Levy, professor of physiology at the Human and Molecular Genetics Center of the Medical College of Wisconsin, studies cystic fibrosis (CF). Although CF is caused by mutations in the CF transmembrane conductance regulator gene, patients with identical genetic mutations in this gene can suffer variable levels of lung disease severity.
Dr. Levy is harnessing a functional genomics technique initially used to determine when people at high risk for type 1 diabetes—such as those who have siblings with the disease—might go on to develop the disease. The technique relies on the observation that inflammatory cytokines are often upregulated in those CF patients who will go on to have more severe pulmonary disease, but at levels too low to detect.
When serum from those patients is cultured with peripheral blood mononuclear cells from healthy individuals, however, it induces an inflammatory signature that can be readily detected, thus distinguishing CF patients at risk for more severe lung disease.
“We have validated this technology and plan to use it to correlate gene-expression status to disease status and pulmonary function,” Dr. Levy said. As progressive pulmonary disease is the major cause of morbidity and mortality in CF patients, identifying markers of such disease should help plan appropriate, individualized courses of treatment.