Figure 1.
Sequencing reads from 2705 individuals (low-coverage whole-genome and exome sequencing) from 26 populations comprising a total of 922 billion reads (87.1 Tbp) used for the 1000GP population BWT. Reads were first error-corrected using a Cortex graph (Iqbal et al. 2012). The error-corrected reads were then trimmed to either 100 or 73 bp, unique sequences identified on the forward strand, quality values discarded, and the metadata stored in a separate database. This resulted in 4.9 Tbp consisting of 53 billion nonredundant reads.
