Efficient genotyping of individuals using overlapping pool sequencing and imputation

Next generation sequencing technologies are rapidly decreasing the cost of obtaining genetic information. The cost for utilizing one of these technologies consists of a sample preparation step and a sequencing step of the prepared sample. The dramatic increase in the efficiency of the sequencing technology makes the costs of the sequencing step negligible for small target regions. Thus the main remaining cost is the sample preparation step. Using overlapping sequencing pools where samples are mixed together into pools which are prepared and sequenced together has been shown to reduce the cost significantly for collecting information on genetic variants which only occur in a few of the samples. These methods utilize ideas from compressed sensing. In this paper, we extend this approach to utilize additional information from reference genetic variation datasets which provide the correlation structure between genetic variants. Utilizing this information, we can significantly increase the efficiency of overlapping pool sequencing.