This paper investigated the effects of fixation duration, block storage duration, extraction method and uracil-N-glycosylase (UNG) treatment on sequencing quality control metrics and deamination events in formalin-fixed paraffin-embedded (FFPE) specimens.
Formalin fixation increases deamination mutation signature but should not lead to false positive mutations in clinical practice.
Prentice LM, Miller RR, Knaggs J, Mazloomian A, Aguirre Hernandez R, Franchini P, Parsa K, Tessier-Cloutier B, Lapuk A, Huntsman D, Schaeffer DF, Sheffield BS
Purpose
Conclusion
Raw read counts were unaffected by up to 48 h of formalin fixation and extraction method (Qiagen FFPE kit, Qiagen GeneRead Kit with UNG, and Qiagen GeneRead Kit withot UNG), but the percentage of mapped reads decreased and the number of deamination single nucleotide variants (SNVs) increased with progressive fixation or block storage. Treatment with UNG decreased the number of deamination SNVs detected in specimens fixed for 48 h or stored as FFPE blocks for 20 years. The highest false positive variant allele frequency (VAF) was found in an UNG-treated specimen (1.48%), however as it was only 1.48% the authors report it is below clinical thresholds.
Studies (3)
Purpose
This study investigated the effects of extraction method on sequencing quality control metrics and deamination events in twenty colorectal specimens. No details of specimen processing were included and the authors state fixation times were unknown. However, all specimens were collected between 2015 and 2016. DNA was extracted from FFPE specimens using (i) the Qiagen FFPE kit after deparaffinization with mineral oil, (ii) the Qiagen GeneRead Kit with UNG, and (iii) the Qiagen Gene-Read Kit without UNG. Adapters were added using the Nextera XT index kit and specimens were sequenced on a MiSeq. SNVs were identified using MutationSeq v4.3.8 software. Deamination counts (C>T, G>A) that occurred at any VAF were compared against those of all other substitutions.
Summary of Findings
Raw read counts and read quality were comparable among case-matched FFPE specimens processed using three different DNA extraction methods. After normalization to the UNG treated specimens, 0-111 deamination events were identified per sample, with no differences observed between specimens extracted using the Qiagen FFPE kit and the Qiagen GeneRead kit.
Biospecimens
TissueColorectalPreservation Types
FormalinDiagnoses
NeoplasticCarcinomaPlatforms
DNANext generation sequencingPre-analytical Factors
Classification Pre-Analytical Factor Value(s) Analyte Extraction and Purification Analyte isolation method Qiagen FFPE kit using mineral oil for deparaffizationQiagen Gene-Read Kit without UNGQiagen GeneRead Kit with UNG
Purpose
This study investigated the effects of the duration of formalin fixation on sequencing quality control metrics and deamination events in three colon adenocarcinoma specimens. Specimens were divided such that one piece was snap-frozen in liquid nitrogen and stored at -80˚C while the remaining pieces were fixed in formalin for 2, 15, 24 or 48 h. Case-matched normal tissue was snap-frozen or fixed in formalin for 2 h. After fixation, specimens were processed and paraffin-embedded. Specimen pathology was confirmed by H&E staining. DNA was extracted from snap-frozen specimens using the Qiagen Gentra Puregene kit. FFPE sections were deparaffinized in mineral oil FFPE specimens prior to DNA extraction using the QIAamp FFPE Tissue kit followed by the Qiagen Gene Read Kit with or without UNG. Adapters were added using the Nextera XT index kit and specimens were sequenced on a MiSeq. SNVs were identified using MutationSeq v4.3.8 software. Deamination counts (C>T, G>A) that occurred at any VAF were compared against those of all other substitutions.
Summary of Findings
Specimens fixed in formalin for different durations were found to have the same grade, histologic subtype, cellularity and tumor content. Library yields as determined by Qubit and bioanalyzer and raw read counts were unaffected by fixation duration, and the authors report electrophoresis showed smears of comparably sized DNA for all FFPE specimens. However, the percentage of mapped reads decreased from 98.5% and 100.2% in specimens fixed for 2 h to 97.2% and 97.8% in specimens fixed for 48 h when DNA was extracted by QIAamp (P=0.02) and GeneRead with UNG (P=0.01). The authors calculated this to be a rate of decline of approximately 0.3% of mapped reads per hour. Deamination SNVs (C>T or G>A) occurred at higher rates than other SNVs in all three specimens fixed for 48 h, but not for shorter durations. Treatment with UNG decreased the number of deamination SNVs detected, such that their incidence was similar to those in specimens fixed for 2 h. Interestingly the highest false positive VAF was found in a UNG-treated specimen (1.48%), however the authors noted it was below the clinical thresholds. Intratumoral heterogeneity was observed in two of three patients, as the same mutations were not identified in all specimens from those patients.
Biospecimens
TissueColorectalPreservation Types
FormalinFrozenDiagnoses
NeoplasticCarcinomaNeoplasticNormal AdjacentPlatforms
DNAAutomated electrophoresis/BioanalyzerDNAElectrophoresisDNAFluorometryDNANext generation sequencingMorphologyH-and-E microscopyPre-analytical Factors
Classification Pre-Analytical Factor Value(s) Biospecimen Preservation Time in fixative 15 h2 h24 h48 hAnalyte Extraction and Purification Analyte isolation method Not treated with UNGTreated with UNG
Purpose
This study investigated the effects of FFPE block age on the number of deamination SNVs detected and NGS quality metric results. FFFPE blocks of normal adjacent and colorectal adenocarcinoma specimens were obtained from three patients each year in 1994, 2004, and 2014. DNA was extracted using the Qiagen QiaAmp FFPE Tissue Kit followed by the Qiagen Gene Read Kit with or without UNG. Adapters were added using the Nextera XT index kit and specimens were sequenced on a MiSeq. SNVs were identified using MutationSeq v4.3.8 software. Deamination counts (C>T, G>A) that occurred at any VAF were compared against those of all other substitutions.
Summary of Findings
The percentage of deamination SNVs increased with progressive FFPE block storage, with a median increase of 15 deamination events over the 20 year period examined. The effect was partially reversed by treatment with UNG prior to sequencing, reducing the median increase to 3 deamination events over the 10 year period examined. The median percentage of mapped reads decreased in UNG-treated and -untreated specimens with FFPE block storage from 98.4% and 97.9%for specimens collected in 2014, respectively, to 98.1% and 97.1% for specimens collected in 1994, respectively.
Biospecimens
TissueColorectalPreservation Types
FormalinDiagnoses
NeoplasticCarcinomaNeoplasticNormal AdjacentPlatforms
DNANext generation sequencingPre-analytical Factors
Classification Pre-Analytical Factor Value(s) Storage Storage duration Collected in 1994Collected in 2004Collected in 2014Analyte Extraction and Purification Analyte isolation method Treated with UNGUntreated