In Nature Biotechnology this week, a team led by researchers at the National Institute of Standards and Technology report on ways to meet the need for highly accurate sets of benchmark genotypes across a genome. Focusing on methods to make high-confidence, single-nucleotide polymorphism, indel, and homozygous reference genotype calls for the pilot genome for the Genome in a Bottle Consortium, they integrated and arbitrated between 14 datasets from five sequencing technologies, seven read mappers, and three variant callers. They identified regions for which no confident genotype call could be made, then classified them by reasons for uncertainty. The calls are publicly available to help with benchmarking of any method.
Meanwhile, in Nature Methods, a pair of scientists from the University of Chicago presents new algorithms in genome-wide efficient mixed model association software for fitting multivariate linear mixed models — used to test associations between single-nucleotide polymorphisms and multiple correlated phenotypes — and computing likelihood ratio tests. The algorithms promise greater computation speed, power, and P-value calibration over current methods.