Fig. 2

Integration workflow of Quartet small variant and structural variant benchmark calls. This workflow depicted the integration process to obtain small variant benchmark calls from 108 original GVCF call sets. Numbers in the boxes represented remaining small variants after each data processing step in the grey dotted boxes. Approximately 6 million small variants were discovered in 27 call sets for each Quartet reference sample. About 1.5 million small variants were removed by the voting process (“ Methods”). We merged the four consensus call sets corresponding to the four Quartet samples, and discarded variants that did not reach agreement across 27 replicates in any Quartet sample. Only Mendelian consistent variants, which were shared by twins and following Mendelian inheritance laws and validated by PMRA and PacBio CCS datasets, were kept as small variant benchmark calls