The overall efficiency of DS is calculated by dividing the number of DCS reads by the number of raw reads, and in our hands it can range from 1 to 10%. We have found that two primary factors determine overall data yield: the peak family size and the SSCS:DCS ratio. The peak family size is determined by the amount of DNA used for the PCR and the fraction of a lane dedicated to a sample, and it should optimally be adjusted to six, as discussed previously (see the ‘Setting tag family size by PCR amplification’ section; Fig. 5a). Should a sample exhibit a peak family size that is smaller than what is desired, then an optimal peak family size can be obtained by resequencing the same sample using the information in Supplementary Table 2 and combining the data with the previous run(s). Importantly, the resequenced sample must be a technical replicate; performing the protocol again on the same biological sample will not work. The SSCS:DCS ratio is the other key factor. A ratio of two is the theoretically ideal and
