PipeConcord
pipeconcord compares bioinformatics pipeline outputs with semantic metrics
instead of byte-for-byte snapshots. Each comparator returns the same
ConcordanceReport shape, so one-off checks and batch regression tests can
consume results consistently.
Rename note: the first alpha release used the package name biocompare. Current
and future releases use pipeconcord to avoid confusion with an unrelated
life-science product directory.
The documentation site is intended for users evaluating the project as a tool for regression testing, pipeline validation, and reproducible scientific comparisons.
Current Comparators
| Type | Inputs | Main metrics |
|---|---|---|
table |
CSV/TSV tables | row overlap, numeric similarity, categorical agreement |
deg |
differential expression tables | significant set overlap, logFC correlation, direction agreement |
counts |
gene-by-sample count matrices | sample correlations, library ratios, zero-pattern overlap |
expression |
TPM/FPKM/CPM matrices | expression profile correlation, distribution similarity, top gene overlap |
bed |
BED3+ intervals | bp Jaccard, interval precision/recall/F1 |
fasta/fastq |
sequence records | record overlap, exact sequence agreement, GC/length similarity |
vcf |
VCF calls | normalized allele Jaccard, genotype concordance, AF correlation, Ti/Tv |
bam_stats |
samtools flagstat / samtools stats text |
alignment-rate, duplicate-rate, and count-ratio similarity |
Common Commands
Install from PyPI:
python3 -m pip install pipeconcord
PyPI package: https://pypi.org/project/pipeconcord/
pipeconcord compare file_a.tsv file_b.tsv --type table --key gene_id
pipeconcord batch manifest.tsv --min-concordance 0.95
pipeconcord compare file_a.bed file_b.bed --type bed --format html --output report.html
Documentation
- Quickstart
- API
- Design
- Writing Comparators
- Regression Testing Tutorial
- RNA-seq Tutorial
- Variant Caller Tutorial
- Peak Caller Tutorial