← All participant guides

Short-read assembly

What this exercise tests

You are given paired-end short reads and a result-sheet template. Assemble each genome, inspect read and assembly quality, identify the organism and explain any failed or questionable result.

Before starting

  1. Read installing bioinformatics tools.
  2. Download the R1 and R2 files and result sheet.
  3. Confirm that every public sample identifier has one read pair.
  4. Keep the downloaded files unchanged.
gzip -t input/*.fastq.gz
ls -lh input/

Recommended route

The recommended worked route uses BactScout for read QC and GHRU-assembly for assembly and post-assembly checks. An equivalent documented bacterial assembly workflow is also acceptable unless the exercise states otherwise.

1. Inspect the reads with BactScout

From a BactScout checkout:

pixi run bactscout qc /absolute/path/to/input \
  -o /absolute/path/to/bactscout_results \
  -t 8

Review final_summary.csv and the per-sample reports. Look for low Q30, adapter or ambiguous-base warnings, inadequate coverage, unexpected taxonomy and mixed-sample evidence. Record how warnings affected the final QC decision.

2. Prepare the GHRU-assembly sample sheet

Create a CSV with absolute input paths. Leave long_reads empty for a short-read-only sample.

sample_id,short_reads1,short_reads2,long_reads,genome_size
sample_01,/data/sample_01_R1.fastq.gz,/data/sample_01_R2.fastq.gz,,5.5m

Sample identifiers and paths must not contain spaces. FASTQ inputs must end in .fq.gz or .fastq.gz.

3. Run the assembly workflow

git clone https://github.com/ghruproject/GHRU-assembly.git
cd GHRU-assembly

nextflow run main.nf \
  --samplesheet /absolute/path/to/samplesheet.csv \
  --outdir /absolute/path/to/assembly_results \
  -resume

The short-read route uses Shovill and includes trimming, assembly assessment, taxonomy and contamination-oriented checks. Save the command and execution report.

Evidence to inspect

  • total assembly length;
  • number of contigs and N50;
  • GC content;
  • read depth or mapped-read proportion;
  • taxonomic classification;
  • completeness and contamination evidence;
  • extremely short or duplicated contigs; and
  • agreement between read-level and assembly-level findings.

N50 is not a correctness score. A contiguous assembly can contain the wrong organism, contamination or a misassembly.

Fill the result sheet

  • sample_name: preserve the supplied public identifier.
  • tax_classification: record the organism supported by your analysis.
  • r1 and r2: retain the correct paired filenames.
  • qc: use the exact vocabulary requested by the template.
  • error: give the main reason for a failed result.
  • notes: record warnings, important parameters and interpretation.

Do not copy an expected species label into tax_classification without checking the evidence.

Common problems

Very small assembly: Check read pairing, coverage, trimming and file paths.

Very large assembly: Investigate contamination, duplicated sequence or unfiltered short contigs.

Many short contigs: Review coverage, read quality, mixed content and repeat structure.

QC tools disagree: Recheck the underlying reports before choosing a final call.

Next: submission and troubleshooting.