A working group of experts assembled by the US National Human Genome Research Institute has published guidelines for studying how variations in genetic sequence may cause disease in humans. Not all variations play a role in disease, and incorrect assessment of variants can have adverse effects, including incorrect treatments for patients. The guidelines apply to germline variants with a <1% minor allele frequency and focus on study design, implication at genetic and variant levels, publication and clinical diagnosis. On the gene level, evidence implicating variants in disease are genetic and experimental. Genetic evidence is the demonstration of gene burden, and experimental evidence consists of data supporting implication in protein interactions, biochemical function, expression, gene disruption, model systems and phenotype rescue. At the variant level, evidence is genetic, informatic and experimental. Genetic evidence is in the association, segregation and population frequency of a variant, and informatic evidence comprises conservation of the site of the variation and the predicted effect of the variation on function. Experimental evidence consists of gene disruption, phenotype recapitulation and phenotype rescue. Source: Nature