Supplementary data: Quantifying sequencing error and effective sequencing depth of liquid biopsy NGS with UMI error correction

<b>Supplementary
Figure 1. Ratio of forward-to-reverse reads is no useful quality criterion
for UMI-libraries after error-correction bioinformatics. </b>The
density plots show the position balance of forward reads to reverse reads at
each genomic position, from targeted sequencing of the <i>PIK3CA</i> coding regions. The position balance was computed with
GenSearchNGS, with 1 being optimal and 0 being worst. (A) Position balance in
the four non-UMI-libraries (TruSeq 776-1, 776-2, 779-1, 779-2). The majority
of genomic positions have a near-optimal read balance. (B) Position balance
in the two UMI-libraries (Oncology 776, 779). Nearly all positions are
covered by error-corrected reads oriented in just one genomic direction. The
conventional quality check of forward/reverse balance is therefore no longer
useful after UMI-based error-correction has been performed.

<b>Supplementary Table 3: Deduplication (removal of
redundant sequences) increases the allele frequency of signal noise (PCR or
sequencing errors) and reduces the effective sequencing depth </b>
<b> </b>
<b>Supplementary Table 2: Allelic frequency of PIK3CA
p.E545K mutation for non-UMI libraries vs UMI libraries and bioinformatic
non-filtering vs filtering, in wild-type standard HD776 (0.00%) and mutated
standard HD779 (0.13%)</b>
<b> </b>
<b>Supplementary Table 1: Bioinformatic tools, versions,
and settings that were used</b>