Fig. 1

Strand-specific errors in base calling. Sequence logos are shown for one stretch of sequence from forward (top) and reverse (bottom) reads. The height of each letter gives the fraction of reads at that position that supported that base. Position 6 (and position 3, to a lesser extent) shows a strand-specific error – the correct base is C, but in the reverse reads a T is the most-evidenced base. When forward and reverse reads are pooled, there appears to be two bases at this position