test datasets

Datasets

rnaseq – Bacterial RNA-seq

RNA-seq data from Bifidobacterium pseudolongum subsp. globosum grown under different conditions (strains, time points, media). The 27 paired-end Illumina NovaSeq 6000 runs produce deep BAMs (~58 million mapped reads each) mapped to a single 2.2 Mb reference genome. This dataset is used to benchmark per-feature read counting against featureCounts.

Source:

• BioProject PRJNA1156008;
• Millar et al., Novel amylase genes enable utilisation of resistant starch by bifidobacteria relevant to early-life microbiome development (bioRxiv 2024.10.09.617373).

---

contigs – Metagenome co-assembly

Faecal metagenomes of great tits and blue tits assembled with MEGAHIT into a co-assembly of 3.59 million contigs (2.37 Gbp). Three samples were mapped to the co-assembly, and subsampled chunks of ~10,600 read pairs each were mapped to produce 12 sparse BAMs — a stress test for per-reference overhead with very many small references. Used to benchmark per-contig coverage and counts against coverm.

Source:

• BioProject PRJEB79635 (ENA);
• Pallen et al. (2025), Faecal metagenomes of great tits and blue tits provide insights into host, diet, pathogens and microbial biodiversity (PubMed 40302838).
• MAG and viral catalogues on Zenodo: 10.5281/zenodo.13643296, 10.5281/zenodo.13643307.

---

prophage – Prophage induction coverage

Ruminococcus gnavus cultures sequenced with and without prophage induction. The ATCC 29149 strain has 24 BAMs (3 biological replicates × 4 technical replicates under control and mitomycin C conditions) mapped to a single 3.55 Mb circular chromosome. The CC55_001C strain has 2 BAMs (control and mitomycin C) mapped to a 3.17 Mb circular contig. Coverage rises several-fold over the induced prophage regions, making this a realistic test for coverage track generation and enriched-region detection.

Source:

• Internal data — not publicly available.

---

tradis – Transposon insertion sequencing

TraDIS-Xpress transposon insertion sequencing of Escherichia coli under colistin stress and control conditions. The 30 single-end NextSeq 500 runs are intended for benchmarking insertion-site counting and 5′-end coverage analysis. The FASTQs and BAMs have not yet been fetched or built.

Source:

• ArrayExpress E-MTAB-11807 (colistin stress) and E-MTAB-11808 (controls);
• Yasir et al. (2022), Genome-Wide Analysis of Innate Susceptibility Mechanisms of Escherichia coli to Colistin (Antibiotics 11(11):1668).