It is now possible to create 'PCR-free' Illumina libraries, in which there is no DNA amplification prior to cluster generation and sequencing. A comparison of libraries prepared with our standard Fisher et al. protocol and a PCR-free protocol (Table S1 in Additional file 3) reveals that the PCR-free libraries lead to less bias across all bias motifs on P. falciparum, E. coli, and R. sphaeroides samples. On human samples, PCR-free library construction produced improved coverage of all motifs except for GC ≥ 75% and G|C ≥ 80%. Additionally, the bad promoters, although improved, were still two-fold undercovered. These results suggest that PCR-free library construction reduces, but does not cure, coverage bias.
