The present invention provides a method for detecting a genetic variant in a region of interest in a DNA sample comprising (i) determining, for a given sequencing platform, sequencing process and sequencing depth, the distribution of the number of reads supporting a genetic variant or plurality of genetic variants expected to be observed in the sequencing results of amplification reactions due to amplification and sequencing error (read count distribution); (ii) based on the read count distribution determined in step (i), establishing a threshold frequency at or above which the genetic variant must be observed in sequencing results of amplification reactions to assign a positive determination for the presence of the genetic variant in a given amplification reaction; (iii) partitioning the DNA sample into a plurality of replicate amplification reactions, so that the mean number of amplifiable template molecules of the region of interest in a replicate amplification reaction is fewer than the reciprocal of the threshold frequency determined in step (ii); (iv) performing the amplification reactions of step (iii) and sequencing the products of amplification reactions, (v) based on step (ii) and the results of step (iv), determining the presence / absence of the genetic variant in each replicate amplification reaction; and (vi) integrating the results of (v) to determine the presence / absence of the genetic variant in the region of interest in the DNA sample.