From Dave Smith,Alnwick, Northumberland, UK
While the points made in “Failure to replicate” are all reasonable, there is a potentially valuable explanation for seemingly unreproducible experiments: an additional factor that hasn’t been controlled for (9 April, p 45).
For example, when researchers were faced with different results from genetically identical mice obtained from different suppliers, patient detective work led them to a . Interestingly, not only is the microbiome of people very variable, but so, too, is
