Introduction Usage Parameters Output Releases

Introduction

This document describes the output produced by the pipeline. Most of the plots are taken from the MultiQC report, which summarises results at the end of the pipeline.

The directories listed below will be created in the results directory after the pipeline has finished. All paths are relative to the top-level results directory.

Pipeline overview

The pipeline is built using Nextflow and processes data using the following steps:

  • Alignments - Alignment of the query genomes to the target genome
  • Dot plots - Visualisation of the alignment of the query genomes to the target genome
  • N regions - Coordinate of the N regions on the query and target genomes
  • MultiQC - Aggregate report describing results and QC from the whole pipeline
  • Pipeline information - Report metrics generated during the workflow execution

Each query genome, is aligned to the target genome, and each alignment is visualised with dot plots. The output file names are constructed by concatenating the target and query sample identifiers with a ___ separator (three underscores), to faciliate re-extraction of the IDs from file names.

Assembly statistics

Output files
  • assemblyscan/
    • *.json contains the statistics collected with the assembly-scan software.

Basic statistics on nucleotide content and contig length are collected for aligned genome for later plotting with MultiQC.

Alignments

Output files
  • alignment/
    • *.train is the alignment parameters computed by last-train (optional)
    • *.train.tsv reports some of the parameters computed by last-train for MultiQC (optional)
    • *.m2m.maf.gz is the many-to-many alignment between target and query genomes. (optional through the --m2m option)
    • *.m2o.maf.gz is the many-to-one alignment: each region of the query genome is aligned at most once, while regions of the target genome can be aligned several times. (optional through the --m2m option)
    • *.o2m.maf.gz is the one-to-many alignment: each region of the target genome is aligned at most once, while regions of the query genome can be aligned several times. (optional through the --m2m option)
    • *.o2o.maf.gz is the one-to-one alignment between the target and query genomes.
    • *.o2o.tsv reports nucleotide percent identity of the one-to-one alignment for MultiQC.
    • *.o2o.matrix.txt reports the substitution matrix and some evolutionary distance indices computed from the one-to-one alignment for MultiQC.
    • For each one-to-one alignment there will be an additional file in a format such as Axt, Chain, GFF or SAM/BAM/CRAM if you used the --export_to parameter. These extra files are always compressed with gzip when their format is text-based. The SAM/BAM/CRAM files are always sorted. Their header features all sequences from the target genome, including the ones that did not align to the query so that alignment files can be merged without disturbing the sort order.
    • The target genome sequence, compressed with bgzip and indexed by samtools is also present when BAM or CRAM files are produced.
    • A multi-query CRAM sequence is present when --multi_cram is used, named like the target genome but with the cram suffix.

Genomes are aligned with lastal after alignment parameters have been determined with last-train. Many-to-many alignments are progressively converted to one-to-one with last-split. In a many-to-one alignment, sequences from the target genome can be matched multiple times. This is for instance useful when aligning a diploid or draft assembly (query) to a primary reference genome (target), or aligning a whole-genome-duplicated query to a target with the ancestral ploidy. Conversely, a one-to-many alignment is useful when expecting multiple matches on the query by the target. Please note also that locally duplicated regions will appear as deletions in the one-to-one alignment, as it will only keep one copy.

To speed up alignment, both strands of the target genome are indexed. This doubles memory usage and may produce output files containing -/+ alignments, which are not supported by some downstream pipelines. To disable this behavior, use --strand forward.

Dot plots

Output files
  • alignment/
    • *.m2m.png (optional)
    • *.m2o.png (optional)
    • *.o2m.png (optional)
    • *.o2o.png (optional)

Dot plots are representing the pairwise genome alignments and produced with the last-dotplot tool. By default, their maximal width is fixed to aproximately 1200 pixels, so that the target genome is always represented at the same scale in all plots. In the one-to-one alignment example below, the hg38 human genome (target) is represented on the horizontal axis and a monkey genopme (Macaca mulatta accession number GCA\_049350105.1) on the vertical axis (query). Regions containing unknown (N) sequences are on pink background. Forward (+/+) alignments are plotted in red and reverse (+/– or –/+) in blue. Target (human) contigs are displayed in their original order. Query contigs (monkey) are reordered and possibly reverse-complemented to diagonalise the plot as much as possible. The names of reverse-complemented contigs are printed in blue.

Example of a dot-plot produced by the pipeline after aligning human and macaque genomes

N regions

Output files
  • cutn/
    • <target>.cuts.bed
    • <sample>.bed

The poly-N regions longer than 9 bases in each genome sequence often indicate contig boundaries in scaffolds. Therefore, we marked them in pale red in the dot-plots. They are detected with theseqtk cutN command and its output (in 3-column BED format) is provided in the cutn directory. Sample IDs are constructed to generate file names, except for the target genome which is always called <target>.cuts.bed to avoid filename collisions.

MultiQC

Output files
  • multiqc/
    • multiqc_report.html: a standalone HTML file that can be viewed in your web browser.
    • multiqc_data/: directory containing parsed statistics from the different tools used in the pipeline.
      • /multiqc_data/multiqc_train.txt: table reporting the alignment parameters chosen by last-train, for each sample.
      • multiqc_data/multiqc_last_o2o.txt: table reporting the nucleotide percent identity in the alignments computed by lastal, for each sample.
    • multiqc_plots/: directory containing static images from the report in various formats.
    • assemblyscan_plot_data: GC content and contig length statistics parsed from assemblyscan for MultiQC with a local module.

MultiQC is a visualization tool that generates a single HTML report summarising all samples in your project. Most of the pipeline QC results are visualised in the report and further statistics are available in the report data directory.

Results generated by MultiQC collate pipeline QC from supported tools e.g. FastQC. The pipeline has special steps which also allow the software versions to be reported in the MultiQC output for future traceability. For more information about how to use MultiQC reports, see http://multiqc.info.

The example MultiQC plots below were generated on this pipeline’s full test dataset, which aligns the hg38 human genome to other primate genomes.

Base content

The pipeline reports the base content of every query genome, like in the example below:

Example of a base content report for primate genomes

Contig length statistics

Contig length statistics can be displayed by MultiQC as violin plots.

Example of a contig length report for primate genomes

Training parameters

Alignment parameters computed by last-train can be displayed by MultiQC as violin plots.

Example of alignment parameters for primate genomes aligned to the human genome

Alignment

Alignment statistics can be displayed by MultiQC as violin plots. There is no standard way to compute nucleotide identity (May A. 2004), therefore the pipeline reports two alternatives, including or excluding gaps from the computation.

Example of alignment statistics for primate genomes aligned to the human genome

Pipeline information

Output files
  • pipeline_info/
    • Reports generated by Nextflow: execution_report.html, execution_timeline.html, execution_trace.txt and pipeline_dag.dot/pipeline_dag.svg.
    • Reports generated by the pipeline: pipeline_report.html, pipeline_report.txt and software_versions.yml. The pipeline_report* files will only be present if the --email / --email_on_fail parameter’s are used when running the pipeline.
    • Reformatted samplesheet files used as input to the pipeline: samplesheet.valid.csv.
    • Parameters used by the pipeline run: params.json.

Nextflow provides excellent functionality for generating various reports relevant to the running and execution of the pipeline. This will allow you to troubleshoot errors with the running of the pipeline, and also provide you with other information such as launch commands, run times and resource usage.

Known limitations and troubleshooting

Fragmented assemblies

The pipeline indexes and aligns assemblies with millions of scaffolds, but:

  • Dot plots show only the 100 largest sequences of each genome (last-dotplot --maxseqs, adjustable with --dotplot_options).
  • Downstream software that loads every sequence name can run out of memory or fail. Keeping the chromosome-level sequences, or the few hundred longest ones, before alignment avoids this and makes plots and statistics easier to interpret.

A query has no output

If last-train finds no alignment between the target and a query, for instance because they are too divergent for the selected seed, the ALIGNMENT_TRAIN process fails with exit status

The pipeline ignores this failure and skips the query instead of stopping. Try a denser seed (--seed YASS or --seed MAM8), or provide alignment parameters with --lastal_params.

Memory

With a human-sized target and 12 threads, peak memory is about 40, 65 and 220 GiB with the RY128, RY4 and YASS seeds. Indexing both strands (--strand both, the default) roughly doubles the memory of the index; --strand forward reduces it, but slows down alignment.

Alignments to the reverse strand of the target

With --strand both, some alignments are reported between the reverse strand of the target and the forward strand of the query. Software that assumes the target is always on the forward strand will mishandle them; use --strand forward in that case.

Reproducibility and number of CPUs

The number of threads does not change the alignments, but it can change the order in which they are written. Outputs are therefore bitwise identical only between runs that give LAST processes the same number of CPUs.

Many-to-many alignments

--m2m adds computation time and can take much disk space. Dot plots of many-to-many alignments of repetitive genomes are slow to draw and hard to read; skip them with --skip_dotplot_m2m.

CRAM version

CRAM files are written in version 3.1, the default of samtools. Some older software reads only version 3.0; see the “Export” section of the Methods of the article, or set the --output-fmt-option version=3.0 argument through ext.args2 of ALIGNMENT_EXP and ext.args of ALIGNMENT_MERGE.

On this page