Computational Systems Biology Methods and Protocols.7z

ed.(R 1 o1,R 2 o1)>ed.(R 1 o,R 2 o)<ed.(R 1 o+1,R 2 o+ 1).

Figure4 shows an example of how AfterQC’s overlapping analysis

works.

Based on overlapping analysis, AfterQC can detect mismatches.

If the mismatched pair has unbalanced quality scores, which means

one base has high-quality score (i.e.,>Q30) and the other has very

low-quality score (i.e.,<Q15), AfterQC can automatically correct

the base with low quality. If the quality scores are not unbalanced,

AfterQC can mask them by changing the bases to N or assigning

zero quality scores to them. Based on the mismatches, AfterQC can

evaluate the sequencing error rate and profile the sequencing

error transform distribution (i.e., how many bases are T but

sequenced as C).

Overlapping analysis can be used for automatic adapter cutting.

In the overlapping analysis process, we get the optimal offsetOfor

the best local alignment of each pair. The overlapping length of this

pair can be directly calculated using the offsetO.IfOis found

negative, the bases outside overlapping region will be considered as

a part of adapter sequences and then be cut automatically.

AfterQC is an open source tool: https://github.com/

OpenGene/AfterQC. It is implemented in Python and C++, with

PyPy support enabled. AfterQC generates a standalone HTML

report for each input, with figures plot by Plotly. A sample report

can be found at:http://opengene.org/AfterQC/report.html.

2.2 Molecular
Barcoding Sequencing
and Its Data Analysis

The potential of NGS deep sequencing for ctDNA was hampered

by systemic errors introduced by PCR and sequencing methods

[27, 28]. Molecular indexing combined with deep sequencing

holds great promise to break the limit imposed by PCR and

sequencing errors and enables the detection of rare and ultra-rare

mutations [29, 30].

Tagging individual templates with molecular barcodes has been

proposed and reported since 2007 [31]. The molecular barcodes or

molecular indexes have been given various names, such as unique

identifiers (UID) [29], unique molecular identifiers (UMI) [32],

primer ID [30], duplex barcodes [33], etc. They are usually

designed as a string of totally random nucleotides (such as

NNNNNNNN), partially degenerate nucleotides (such as

Fig. 4How AfterQC’s overlapping analysis works. The edit distance of the overlapped subsequences is 1. A
mismatch pair is found with a high-quality baseAand a very low-quality baseT. ThisTwill be recognized as
wrongly represented and can be corrected

78 Shifu Chen et al.