Table 1.
Various options for PCR-based enrichment and their application in oncology.
Fig 1.
Steps in amplicon enrichment sample preparation with subsequent sequencing on the Illumina and MGI platform.
A – common part, B and C – steps of the Illumina- and DNBSEQ-adapted protocols, respectively. DNB – DNA nanoballs.
Fig 2.
Analysis of biases in amplicon coverage across samples sequenced with diverse Ampliseq-based protocols (DDD/IDI/III – corresponds to variations of PCR mix1 (I/D), ligase/buffer (I/D) and PCR mix2 (I/D); ‘-E’ stands for alternative Digest Mix).
(A) – heatmap of amplicon coverage (% of average coverage) across samples demonstrates significant variation in per-sample set of dropped out (10% of average sample coverage) amplicons depending on i) PCR mix 2 used ii) Digest Mix used. Coverage of amplicons by amplicon GC content analysis in samples sequenced on MGI platform (B) demonstrates extremely high amplitude between coverage of GC-low (30% and lower) and GC-moderate (40% and over) amplicons, while GC-high (70% and higher) amplicons demonstrate only slight decrease in coverage.(C) Per-amplicon coverage correlation between different protocols adopted for Illumina/MGI sequencing platforms.
Fig 3.
In silico sensitivity for the detection of pathogenic variants in the BRCA1/2 genes (ABC plus panel) and predictive biomarkers of ESMO Scale for Clinical Actionability of Molecular Targets (ESCAT) level I across major oncogenes (Atlas plus panel) was determined using EphaGen software.
The analysis was conducted on datasets comprising 29 samples sequenced using the DNBSEQ-adapted protocol and compared with results of blood whole exome sequencing (WES) involving a total of 13 samples.
Fig 4.
Characteristics of DNA libraries sequenced using MGI WES protocol.
Sequence context of DNA fragment ends demonstrates equivalence of 5’-3’/3’-5’ DNA strands sequencing in terms of biased representation of fragments with terminal G or C nucleotides (A). The bias was consistent across all samples (B) GC-content analysis of sequenced DNA fragments in comparison with expected by random sampling demonstrates slight overrepresentation of GC moderate (40-60%) regions at the expense of GC low (<30%)/high (>70%) (C) along with high variation in GC-low/moderate/high fragments representation across 13 samples (D). Fragment distribution by length across 13 samples (F).
Fig 5.
Characteristics of Atlas plus and ABC plus panel designs and DNA fragments sequenced using DNBSEQ-adapted protocol.
Sequence context expected by panel (amplicon) design (A, B) and sequenced (C, D) DNA fragment ends demonstrates absence of nucleotide composition based bias in fragments representation (B). GC content analysis of sequenced DNA fragments in comparison with expected by panel design (E, F) demonstrates significant overrepresentation of GC moderate (35-55%) amplicons at the expense of GC low (<30%)/high (>70%) amplicons. (F) – fragment representation by length across 29 samples.