Table 2 presents relative coverage of the six motifs across 16 data sets. From data set 14 (120-fold coverage of Illumina from the human genome) we also chose ten 0.5-fold random subsets, in each case computing the relative coverage across the motifs. For each motif, we show (data set 14') the mean of these ten measurements, which for all motifs were within 0.01 of the full sample value, and the observed standard deviation, which for all motifs was approximately 0.02 or less. This shows that for the human genome, relative coverage of the six motifs can be accurately assayed using low coverage.
