nf-core pipeline · adna, ancient-dna, coprolite
nf-core/coproid
Identifies the biological host of ancient and modern faeces using endogenous DNA and microbiome machine learning.
What nf-core/coproid is
nf-core/coproid is a bioinformatics analysis pipeline designed to identify the biological origin or maker of Illumina-sequenced faeces and palaeofaeces (coprolites). Determining the host species of archaeological faecal remains is frequently challenging due to degraded endogenous material, dietary background, and environmental contamination.
Originally developed by Maxime Borry and Meriam Van Os, the pipeline resolves host identity by synthesising two distinct lines of evidence: comparative mapping of candidate host ancient DNA and machine-learning classification of the sample's gut microbial community against modern reference datasets.
What the pipeline does
The workflow begins by performing quality control on single-end or paired-end Illumina reads using FastQC, followed by adapter trimming and low-complexity read removal using fastp. Filtered reads are mapped against two or three candidate host genomes using Bowtie2. Alignments undergo Lowest Common Ancestor (LCA) analysis with sam2lca to filter out non-specific alignments and retain only genome-specific reads. To verify the authenticity of ancient host reads, DNA damage patterns and deamination profiles are calculated using pyDamage and DamageProfiler, yielding a normalised host DNA species ratio.
In parallel, reads that do not map to the candidate host genomes are taxonomically profiled using Kraken 2. SourcePredict then applies machine learning to these taxonomic profiles to evaluate microbial resemblance to reference host gut metagenomes. Finally, the pipeline integrates the host DNA proportion and the microbiome prediction proportion to determine the most probable host, collating results across samples using MultiQC and an automated Quarto report.
Key facts
- Workflow engine
- Nextflow DSL2
- Primary developers
- Maxime Borry and Meriam Van Os
- Sequencing data inputs
- Single-end or paired-end Illumina FASTQ reads
- Host mapping tools
- Bowtie2 and sam2lca
- Damage estimation tools
- pyDamage and DamageProfiler
- Microbiome profiling tools
- Kraken 2 and SourcePredict
- Reporting tools
- MultiQC and Quarto
What you can use it for
- Coprolite host identification
- Distinguishing whether archaeological faecal specimens originated from humans or other candidate animals by combining endogenous host read counts and gut taxonomic profiles.
- Ancient DNA damage assessment
- Evaluating post-mortem cytosine-to-thymine deamination and fragmentation characteristics in recovered host reads using pyDamage and DamageProfiler to ensure authenticity.
- Faecal microbiome source tracking
- Profiling unmapped gut microbial sequences using Kraken 2 and classifying them against modern reference metagenomes with SourcePredict.
Run it here
The research agent runs nf-core/coproid with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
Strengths and limits
Good at
- Integrates both endogenous host DNA quantification and metagenomic microbiome classification to assign host origin.
- Incorporates dedicated ancient DNA damage estimation tools, specifically pyDamage and DamageProfiler.
- Applies Lowest Common Ancestor filtering with sam2lca to eliminate reads aligning ambiguously across target genomes.
- Generates unified multi-tool summaries and sample-level reports through MultiQC and Quarto.
Watch out for
- Candidate reference genomes must be sourced from NCBI with valid taxonomy identifiers for sam2lca compatibility.
- Requires external pre-built databases for Kraken 2 as well as source and label reference files for SourcePredict.
- Comparative host mapping is designed to evaluate between two or three target genomes.
Questions
Can I run nf-core/coproid without my own cluster?
The research agent runs nf-core/coproid with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
What licence is nf-core/coproid under?
MIT, like every nf-core pipeline.
What inputs are required to run the pipeline?
The pipeline requires a sample sheet of single- or paired-end FASTQ files, a genome sheet specifying two or three NCBI reference genomes, a Kraken 2 database, and source and label files for SourcePredict.
Why must reference genomes come from NCBI?
The sam2lca lowest common ancestor tool extracts taxonomy identifiers (taxids) directly from NCBI sequence headers to differentiate between candidate species.
How can the workflow run in an offline or restricted HPC environment?
Users can pre-download the sam2lca database and pass its location directly via the --sam2lca_db and --sam2lca_acc2tax parameters to avoid automatic download failures.
How does the pipeline combine host and microbial evidence?
It calculates a normalised host DNA proportion from genome-specific mapping and combines it with the machine-learning proportion estimated by SourcePredict from Kraken 2 taxonomic profiles.
How to cite it
doi: 10.5281/zenodo.2653756
Written from nf-core/coproid's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.