SEQC2 ctDNA Reference Samples (Liquid-Biopsy Proficiency Study)

A public reference dataset for benchmarking the detection of circulating tumor DNA (ctDNA) at the low variant allele frequencies that define minimal residual disease (MRD). Produced by the FDA-led SEQC2 consortium (Oncopanel Sequencing Working Group), it uses contrived plasma-mimicking samples built from two well-characterized reference materials: Sample A, which carries roughly 40,000 known variants, and Sample B, the matched background. Sample A was diluted into Sample B at 1:4, 1:24, and 1:124 to create Samples D, E, and F, producing a descending allele-frequency ladder that simulates ctDNA at progressively lower abundance. Twelve independent laboratories each ran four technical replicates per ctDNA sample, yielding 359 sequenced DNA libraries across Illumina NextSeq 500, Illumina NovaSeq 6000, and Thermo Fisher Ion S5 XL. Because the truth set is known, the data supports direct measurement of assay sensitivity, limit of detection, cross-lab and cross-platform reproducibility, and false-positive rates, the core analytical-validity questions for any MRD assay. The S3 files mounted here are a curated subset: 10 representative libraries (about 28 GB) covering the full A/B/D/E/F dilution ladder across Ion Torrent, NextSeq, and NovaSeq, plus the run sheet. The complete 359-library collection remains on the NCBI Sequence Read Archive under accession SRP296025 (Gong et al., Scientific Data, 2022).

Category
Genomics & Bioinformatics
Provider
FDA SEQC2 Consortium (Oncopanel Sequencing Working Group)
License
Public (FDA SEQC2 consortium data, NCBI SRA)
DOI
10.1038/s41597-022-01276-8
Hosted on
Amazon S3, bucket scigantic-mrd-liquid-biopsy/seqc2-ctdna-proficiency/, us-east-1
Access
Free and public. No account with the data provider is needed.
Source
https://www.ncbi.nlm.nih.gov/sra/?term=SRP296025
Data structure and loading code

Fields

Runstring (SRR accession — the unit you pre
Sample Namestring (A | B | D | E | F — the dilution
Platform / Instrumentstring (Illumina NextSeq 500 | NovaSeq 6
center_namestring (one of the 12 contributing labor
LibraryName / replicatestring (1 of 4 technical replicates per
Bases / Bytesnumber (run size — ultra-deep targeted p

Load the data

# api archive: no mount. Pull the SRP296025 run table first, then fetch only what you need.
# pip install pysradb
from pysradb import SRAweb
df = SRAweb().sra_metadata("SRP296025", detailed=True)
print(len(df), "runs")
# tabulate the dilution ladder x platform so you know which runs are the low-VAF (Sample F) tier
cols = [c for c in df.columns if c.lower() in ('library_name','sample_title','instrument_model','platform','center_name')]
print(df[cols].head(20))

Written for a Scigantic notebook, where /mnt/archive is the mounted dataset.

Analyze this dataset without downloading it

On Scigantic, SEQC2 ctDNA Reference Samples (Liquid-Biopsy Proficiency Study) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

Related Genomics & Bioinformatics datasets