Library construction methods affect evenness of coverage [20-22]. Table 2 includes human Illumina data produced using the methods described in Aird et al. [20] that are illustrative of this, showing a striking improvement at high GC when the PCR enzyme Phusion HF (data set 10) was supplemented by betaine (data set 11) or replaced by AccuPrime Taq HiFi (data set 12). Figure 3 shows a marked flattening of relative coverage between 15% and 70% GC when we replaced some reagents in the low-input Fisher et al. protocol (data set 13) [31] with reagents from Kapa Biosystems (data set 14) (see Materials and methods), although the large improvement at low-GC was partly offset by a small decline in high-GC coverage (Figure 3, Table 2). Oyola et al. [21] achieved a similar improvement in low-GC coverage of P. falciparum by utilizing Kapa HiFi enzymes and the PCR additive tetramethylammonium chloride in library construction.
