In the early, online version of the Proceedings of the National Academy of Sciences, researchers from China and the US describe DNA methylation profiles that appear to show promise for distinguishing tumor tissue from neighboring normal tissue, along with potential markers for diagnosing cancer and predicting metastases. The team turned to genome-wide methylation data for more than 1,600 breast, colon, liver, and lung cancer samples profiled from the Cancer Genome Atlas effort, which they analyzed alongside 173 nearby matched normal samples with a machine learning method. The search led to candidate methylation markers that were subsequently assessed in another 1,600 tumor or matched normal samples from TCGA or a Chinese cancer cohort, leading to a methylation signatures and related expression patterns that appeared to inform diagnoses or metastases.
Investigators from centers in the US and Switzerland explore interactions between selection and gene flow in the primarily out-crossing yellow monkeyflower Mimulus guttatus and a sister species called M. nasutus, which is capable of self-fertilization. Using information from DNA polymorphisms in the plant genomes and the recombination rate variation patterns across the genome, the team attempted to retrace selection strength and timing in relation to gene flow. With whole-genome population-pooled sequence data for hundreds of M. guttatus individuals in two California populations growing with or without serpentine soil, for example, the authors saw adaptive divergence signals influencing gene flow between the two populations in ways that promoted the maintenance of local adaptations.
Finally, a National Center for Biotechnology Information-led team takes a look at prokaryotic gene dynamics stemming from selection or intrinsic loss bias using a mathematical model that considers genome evolution, comparative genomics, mutation accumulation profiles, and computer simulations. Based on analyses of gene family patterns, gene duplications and losses, and other information gleaned from almost 700 bacterial and archaeal genomes, the authors conclude that "genes involved in key informational and metabolic pathways are subject to strong selection, whereas most of the others are effectively neutral or even burdensome."