In a paper published online in advance in Genome Biology this week, a trio of investigators at the University of Michigan discusses how piRNAs promote genome surveillance in Caenorhabditis elegans "by triggering siRNA-mediated silencing of non-self DNA in competition with licensing programs that support endogenous gene expression."
Over in BMC Genomics, researchers at the Wellcome Trust Sanger Institute compare the performance of Ion Torrent's PGM, Pacific Biosciences' RS, and Illumina's MiSeq, sequencing a set of four microbial genomes on each platform. The team reports that sequences generated by all three platforms were usable, though there are "key differences between the quality of that data and the applications it [each sequencer] will support." Among them, the researchers say, is that they could "call slightly more variants from Ion Torrent data compared to MiSeq data, but at the expense of a higher false positive rate," and that "variant calling from Pacific Biosciences data was possible but higher coverage depth was required."
Harvard Medical School's Marco Ramoni and his colleagues ask: "How accurate can genetic predictions be?" They address the question using mathematics — deriving the absolute limits that the heritability and prevalence of the tested trait impose on accuracy "in the absence of any distributional assumptions on risk," Ramoni et al. write. The researchers present the limits they ascertained "in terms of the best-case receiver-operating characteristic curve, consisting of the best-case test sensitivities and specificities, and the AUC measure of accuracy," and apply them to genetic predictions of type 2 diabetes and breast cancer.
Elsewhere in BMC Genomics, the University of Melbourne's Kathryn Holt and her colleagues present SRST, a software tool "for quick and accurate retrieval of sequence types from short-read sets, using inputs easily downloaded from public databases." Among potential applications, SRST may be useful for "quality control for high-throughput sequencing projects, plasmid MLST [multilocus sequence typing], and analysis of genomic data during outbreak investigation," Holt et al. write.