sequencing process

We generated at least 6 human genome coverage of sequence per week-long run. Each full run consisted of 50 channels distributed across two flow cells. We combined data from four instrument runs, during which 172 of the 200 channels were loaded with P0 genomic DNA. Sequence data were mapped to the NCBI 36 reference human genome (hg18) using the open-source aligner IndexDP; 63% of the raw reads were aligned (Fig. 1a), yielding a total useful coverage of 28.