In PLoS Computational Biology this week, the University of British Columbia's Kelvin Zhang and the Ontario Institute for Cancer Research's Francis Ouellette present an approach to predict cancer outcomes based on relationships "between protein structural information, protein networks, gene expression data, and mutation data," which together form the "domain interaction network in human proteome." The team's approach first scored each protein "by quantifying the domain connections to its interacting partners and the somatic mutations present in the domain," it writes. From there, Zhang and Ouellette add, proteins whose scores surpass a preset threshold can be effectively defined as gene signatures for prognosis.
A trio of investigators at the Indian Institute of Science in Bangalore this week reports a 10-microRNA signature that it says is predictive of glioblastoma patient survival. Using miRNA expression information derived from The Cancer Genome Atlas glioblastoma data set, the team applied Cox regression analysis on a training set of patients chosen at random to formulate risk scores based on median survival times. Of the 10 predictive miRNAs they identified, "seven were found to be risky miRNAs and three were found to be protective," the authors write in PLoS One. The team validated its 10-miRNA signature in a testing set.
In another PLoS One paper, a team led by researchers at the University of California, Irvine, confirm an association between a previously described polymorphism in the transferrin gene and iron status. In a genome-wide association study on 336 iron-deficient cases and 343 controls, the Irvine-led team found five SNPs significantly associated with at least one iron measure; of those, the researchers replicated the association of "rs3811647 on [chromosome] 3q22, a … SNP in the transferrin gene region" with total iron binding capacity in a set of 71 cases and 161 controls.
Over in PLoS Genetics, a team led by investigators at Johns Hopkins University shows that an "ancestral mutation in telomerase causes defects in repeat addition processivity and manifests as familial pulmonary fibrosis." The team "identified two pulmonary fibrosis families that share two non-synonymous substitutions in the catalytic domain of the telomerase reverse transcriptase gene hTERT: V791I and V867M," it reports. In their in vitro functional studies, the researchers found that the double mutant showed "severe defects in telomere repeat addition processivity."