Skip to content

usage.md

Latest commit

 

History

History
381 lines (277 loc) · 30.9 KB

usage.md

File metadata and controls

381 lines (277 loc) · 30.9 KB
Raw

IntGenomicsLab/lrsinglecell: Usage

Documentation of pipeline parameters is generated automatically from the pipeline schema and can no longer be found in markdown files.

Samplesheet input

You will need to create a samplesheet with information about the samples you would like to analyse before running the pipeline. Use this parameter to specify its location. It has to be a comma-separated file with 4 columns, and a header row as shown in the examples below.

--input '[path to samplesheet file]'

Full samplesheet

The example samplesheet.csv below contains a single FASTQ file per biological replicate with sample specific cell counts.

sample,fastq,cell_count,type
CONTROL_REP1,AEG588A1_S1.fastq.gz,5000,cdna
CONTROL_REP2,AEG588A2_S1.fastq.gz,6000,cdna
TREATMENT_REP1,AEG588A4_S1.fastq.gz,5500,cdna
TREATMENT_REP2,AEG588A5_S1.fastq.gz,6000,cdna
CONTROL_REP1,AEG588A1_S1.fastq.gz,5000,dna
CONTROL_REP2,AEG588A2_S1.fastq.gz,6000,dna
TREATMENT_REP1,AEG588A4_S1.fastq.gz,5500,dna
TREATMENT_REP2,AEG588A5_S1.fastq.gz,6000,dna
Column Description
sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. Spaces in sample names are automatically converted to underscores (_).
fastq Full path to FastQ file for Oxford Nanopore. File has to be gzipped and have the extension ".fastq.gz" or ".fq.gz".
cell_count Expected number of cells/nuclei. This value is used by the barcode calling tool (BLAZE and/or Flexiplex) as a baseline when determining an acceptable number of detected barcodes.
type An optional column specifiying whether the sample is DNA or cDNA. If omitted, the default cdna is used.

Note: DNA samples are only compatible with flexiplex demultiplexing.

An example samplesheet has been provided with the pipeline.

Multiple runs of the same sample

The sample identifiers have to be the same when you have re-sequenced the same sample more than once e.g. to increase sequencing depth. The pipeline will concatenate the raw reads before performing any downstream analysis. Below is an example for the same sample sequenced across replicates 1 and 4 (REP1 and REP4 respectively):

sample,fastq,cell_count,type
CONTROL_REP1,AEG588A1_S1.fastq.gz,5000,cdna
CONTROL_REP1,AEG588A1_S2.fastq.gz,5000,cdna
CONTROL_REP2,AEG588A2_S1.fastq.gz,2000,cdna
CONTROL_REP3,AEG588A3_S1.fastq.gz,7500,cdna
CONTROL_REP4,AEG588A4_S1.fastq.gz,9000,cdna
CONTROL_REP4,AEG588A4_S2.fastq.gz,9000,cdna
CONTROL_REP4,AEG588A4_S3.fastq.gz,9000,cdna

Running the pipeline

The typical command for running the pipeline is as follows:

nextflow run IntGenomicsLab/lrsinglecell \
  --input ./samplesheet.csv \
  --outdir ./results \
  --genome_fasta /path/to/genome.fa \
  --transcript_fasta /path/to/transcriptome.fa \
  --gtf /path/to/file.gtf \
  --quantifier "isoquant,oarfish" \
  --demux_tool_cdna flexiplex \
  --demux_tool_dna flexiplex \
  --barcode_format 10X_3v3 \
  -profile <docker/singularity/institute>

Please note that while the above command specifies both transcriptome and genome fasta files, only one is needed for the pipeline and is dependent on which quantifier you wish to use. Isoquant requires a genome fasta, while oarfish requires a transcript fasta. Furthermore, if you have any DNA samples, the genome_fasta is required. Additionally, for the quantifier parameter in the above command, we've listed the quantifiers as a comma-delimited string. It is possible to only use one quantifier, and can be accomplished by just providing the name of the quantifying tool you wish to run as a single value, i.e. providing oarfish if you only wish to run oarfish.

Quantifiers

Three quantifiers are available through --quantifier, and any combination of them can be requested as a comma-delimited string:

Quantifier Reference required Levels Notes
isoquant genome_fasta + gtf gene + transcript Quantifies from a genome alignment
oarfish transcript_fasta transcript Quantifies from a transcriptome alignment; always deduplicates
lrkallisto genome_fasta + gtf gene + transcript Pseudoaligns straight from the FASTQ; quantification uses no alignment

--quantifier only applies to cdna samples. It is required whenever the samplesheet contains at least one cdna row, and can be omitted entirely for a DNA-only run: DNA samples are aligned and deduplicated but never quantified. If a DNA-only samplesheet is given without --quantifier, the pipeline runs the DNA path only; if a cdna row is present without --quantifier, the run stops with an error.

lr-kallisto

lr-kallisto is the long-read mode of kallisto (kallisto --long). Unlike the other two quantifiers, it does not align reads: it pseudoaligns them against a k=63 index and quantifies with a long-read expectation-maximisation step. That makes the quantification itself considerably faster.

The cDNA genome alignment is produced regardless of quantifier, so a BAM and its alignment-derived QC are still available under <sample>/cdna/genome/ even on an lr-kallisto-only run. lr-kallisto does not read it. If you want lr-kallisto's speed and do not need the BAM, --skip_cdna_dedup drops the expensive umi_tools chain and publishes the mapped-only BAM instead.

Because lr-kallisto locates the cell barcode and UMI positionally rather than by name, the pipeline moves them out of the flexiplex read name into a synthetic barcode read before handing the FASTQs to kallisto bus. This reuses the barcodes that flexiplex already corrected against the whitelist, so cell barcodes are directly comparable with the other quantifiers. UMI deduplication is performed by bustools while counting, so --dedup_tool and --skip_dedup have no effect on this path. The published gene/ matrix is an integer UMI count matrix (bustools count --genecounts) and can be given directly to CellBender or any other tool that expects counts; the transcript/ matrix and the gene abundances under quant/ are kallisto quant-tcc EM estimates and are fractional. See docs/output.md for the difference.

This means lr-kallisto has some extra requirements:

  • --genome_fasta and --gtf must both be provided. The transcript sequences are extracted from them with gffread, the transcript-to-gene mapping is derived from the same GTF, and the genome is used as the kallisto d-list so that reads originating outside the transcriptome are discarded rather than misassigned. Deriving both from one GTF is what guarantees the index sequence names and the mapping agree.
  • --demux_tool_cdna must be flexiplex; the blaze path keeps barcodes in a separate per-read table rather than in the read name.
  • --custom_flexiplex_barcode_cdna cannot be used, because the barcode and UMI widths are needed in order to locate them. Use one of the --barcode_format presets instead.

The k-mer length, platform and d-list can be tuned with --kallisto_kmer_size (default 63), --kallisto_platform (ONT or PacBio, default ONT) and --kallisto_dlist (default true). --kallisto_threshold (default 0.8) is also accepted, but it is inert — see the last warning in this section.

Warning

To turn the d-list off, write --kallisto_dlist=false with an equals sign. params.kallisto_dlist defaults to a boolean, so Nextflow treats the space-separated --kallisto_dlist false as a bare flag and drops the false as a positional argument: the d-list stays on and the only sign is a positional argument warning in the log. The same applies to any other boolean parameter you want to set to false on the command line.

Important

Check p_pseudoaligned in bus/run_info.json after your first run. --kallisto_dlist costs real sensitivity; --kallisto_kmer_size is already at its best value:

--kallisto_dlist=false --kallisto_dlist=true (default)
--kallisto_kmer_size 31 31.5% 18.8%
--kallisto_kmer_size 63 (default) 42.8% 30.1%

Measured on 9,805,796 flexiplex-demultiplexed ONT reads against GRCh38 (CellRanger refdata-gex-GRCh38-2020-A). On the same reads minimap2 places 95% and isoquant then assigns 45.8% of them to a gene, so 42.8% is in the same range as the alignment-based path rather than an order of magnitude below it.

  • k-mer length. Pseudoalignment needs exact k-mer matches, so k interacts with per-base error rate — but k=63 beats k=31 on real data at both d-list settings, and it wins by most on the reads that matter: of the reads isoquant assigns confidently to a single gene, k=63 recovers 88.6% against k=31's 58.3% (d-list off). Do not lower k to chase a mapping rate. (An earlier version of this table reported the opposite ordering from the chr21 test_lrkallisto profile. That reference is too small for the comparison to carry: every rate there was under 21%, and the k ordering inverts on a whole transcriptome.)
  • d-list. Using the genome as a d-list discards reads containing k-mers that are in the genome but not the transcriptome, which is what stops intronic and genomic reads being force-assigned to a transcript. This is far more punitive for long reads than for short ones: a long read only has to contain a single genome-only k-mer to be rejected. Turning it off recovered 1,253,671 extra reads. Classifying each of them by what isoquant made of the same read: 59% are reads isoquant also assigns to a gene, 22% are intronic reads isoquant places inside a gene but does not count, 14% are reads minimap2 never placed at all, and only 3% are intergenic — the one class that is unambiguously spurious. On this data the d-list costs far more sensitivity than the specificity it buys, so --kallisto_dlist=false is worth trying for ONT cDNA. The pipeline default is still true, which is what lr-kallisto upstream recommends; this was one flow cell of one chemistry, so it is a reason to test the setting on your own data rather than to assume the default is wrong.
  • read length is not the limiting factor. Only 1.1% of these reads were shorter than k=63, so reads too short to carry a single k-mer explain almost none of the loss.

--kallisto_threshold will not recover anything: it is inert.

The defaults are deliberately the upstream-recommended, higher-specificity ones. If your mapping rate is far below the rate you get from minimap2 on the same data, compare against another quantifier before trusting the counts.

Warning

--kallisto_threshold currently has no effect. kallisto bus --long accepts --threshold and silently ignores it: sweeping it over 0.0, 0.1, 0.2, 0.4, 0.6, 0.8 and 0.9 on 1,000,000 reads returned the identical set of 427,251 pseudoaligned reads and the identical resolved transcript set for every one of them — only the equivalence-class numbering moved. --threshold appears in no other kallisto subcommand, and the pipeline quantifies with kallisto quant-tcc --long, so the parameter cannot take effect anywhere on this path. It is retained only so that existing configs keep parsing; do not use it to tune sensitivity.

Reads are pseudoaligned as non-strand-specific, because after flexiplex has trimmed the adapter, barcode, UMI and poly-T the remaining cDNA is not in a consistent orientation. params.stranded therefore does not apply to this path. If you do want to force a strand, override the KALLISTO_BUS arguments with a custom config:

process {
    withName: '.*:QUANTIFY_SCRNA_LRKALLISTO:KALLISTO_BUS' {
        ext.args = { "--threshold ${params.kallisto_threshold} --fr-stranded" }
    }
}

Note

Building the index is cheaper than you might expect for a k=63 index with a whole-genome d-list. Measured on GRCh38 (CellRanger refdata-gex-GRCh38-2020-A, 199,138 transcripts, 345 MB of transcript sequence) with the 3.15 Gb genome as the d-list: 463 s wall on 24 cpus and 3.6 GB peak RSS, producing a 222 MB index. The process_high default is comfortably sufficient — there is no need to route KALLISTO_INDEX to a high-memory queue. If you use a much larger reference, check peak_rss for that process in the execution trace before assuming these numbers carry over.

The pipeline supports barcode identification and extraction through both flexiplex and blaze and can be set through demux_tool_dna (only works with flexiplex for now) and demux_tool_cdna parameters. The barcode format can be specified through the barcode_format parameter. When working with completely custom barcode structures, you can additionally specify these with custom_flexiplex_barcode_dna and custom_flexiplex_barcode_cdna parameters. Note: ensure that you are using flexiplex as the barcode calling tool. This can be a string formatted as follows "-x CTACACGACGCTCTTCCGATCT -b ???????????????? -u ?????????? -x TTTCTTATATGGG -f 8 -e 2", for more information check the documentation: https://davidsongroup.github.io/flexiplex/

Note that the pipeline will create the following files in your working directory:

work                # Directory containing the nextflow working files
<OUTDIR>            # Finished results in specified location (defined with --outdir)
.nextflow_log       # Log file from Nextflow
# Other nextflow hidden files, eg. history of pipeline runs and old logs.

If you wish to repeatedly use the same parameters for multiple runs, rather than specifying each flag in the command, you can specify these in a params file.

Pipeline settings can be provided in a yaml or json file via -params-file <file>.

Warning

Do not use -c <file> to specify parameters as this will result in errors. Custom config files specified with -c must only be used for tuning process resource specifications, other infrastructural tweaks (such as output directories), or module arguments (args).

The above pipeline run specified with a params file in yaml format:

nextflow run IntGenomicsLab/lrsinglecell -profile <docker/singularity/institute> -params-file params.yaml

with

input: "./samplesheet.csv"
outdir: "./results/"
genome_fasta: "/path/to/genome.fa"
transcript_fasta: "/path/to/transcript.fa"
gtf: "/path/to/file.gtf"
quantifier: "isoquant|oarfish|lrkallisto|isoquant,oarfish"
barcode_format: "10X_3v3"
<...>

You can also generate such YAML/JSON files via nf-core/launch.

Updating the pipeline

When you run the above command, Nextflow automatically pulls the pipeline code from GitHub and stores it as a cached version. When running the pipeline after this, it will always use the cached version if available - even if the pipeline has been updated since. To make sure that you're running the latest version of the pipeline, make sure that you regularly update the cached version of the pipeline:

nextflow pull IntGenomicsLab/lrsinglecell

Reproducibility

It is a good idea to specify the pipeline version when running the pipeline on your data. This ensures that a specific version of the pipeline code and software are used when you run your pipeline. If you keep using the same tag, you'll be running the same version of the pipeline, even if there have been changes to the code since.

First, go to the IntGenomicsLab/lrsinglecell releases page and find the latest pipeline version - numeric only (eg. 1.3.1). Then specify this when running the pipeline with -r (one hyphen) - eg. -r 1.3.1. Of course, you can switch to another version by changing the number after the -r flag.

This version number will be logged in reports when you run the pipeline, so that you'll know what you used when you look back in the future. For example, at the bottom of the MultiQC reports.

To further assist in reproducibility, you can use share and reuse parameter files to repeat pipeline runs with the same settings without having to write out a command with every single parameter.

Tip

If you wish to share such profile (such as upload as supplementary material for academic publications), make sure to NOT include cluster specific paths to files, nor institutional specific profiles.

Core Nextflow arguments

Note

These options are part of Nextflow and use a single hyphen (pipeline parameters use a double-hyphen)

-profile

Use this parameter to choose a configuration profile. Profiles can give configuration presets for different compute environments.

Several generic profiles are bundled with the pipeline which instruct the pipeline to use software packaged using different methods (Docker, Singularity, Podman, Shifter, Charliecloud, Apptainer, Conda) - see below.

Important

We highly recommend the use of Docker or Singularity containers for full pipeline reproducibility, however when this is not possible, Conda is also supported.

The pipeline also dynamically loads configurations from https://github.com/nf-core/configs when it runs, making multiple config profiles for various institutional clusters available at run time. For more information and to check if your system is supported, please see the nf-core/configs documentation.

Note that multiple profiles can be loaded, for example: -profile test,docker - the order of arguments is important! They are loaded in sequence, so later profiles can overwrite earlier profiles.

If -profile is not specified, the pipeline will run locally and expect all software to be installed and available on the PATH. This is not recommended, since it can lead to different results on different machines dependent on the computer environment.

  • test
    • A profile with a complete configuration for automated testing
    • Includes links to test data so needs no other parameters
  • docker
    • A generic configuration profile to be used with Docker
  • singularity
    • A generic configuration profile to be used with Singularity
  • podman
    • A generic configuration profile to be used with Podman
  • shifter
    • A generic configuration profile to be used with Shifter
  • charliecloud
    • A generic configuration profile to be used with Charliecloud
  • apptainer
    • A generic configuration profile to be used with Apptainer
  • wave
    • A generic configuration profile to enable Wave containers. Use together with one of the above (requires Nextflow 24.03.0-edge or later).
  • conda
    • A generic configuration profile to be used with Conda. Please only use Conda as a last resort i.e. when it's not possible to run the pipeline with Docker, Singularity, Podman, Shifter, Charliecloud, or Apptainer.

-resume

Specify this when restarting a pipeline. Nextflow will use cached results from any pipeline steps where the inputs are the same, continuing from where it got to previously. For input to be considered the same, not only the names must be identical but the files' contents as well. For more info about this parameter, see this blog post.

You can also supply a run name to resume a specific run: -resume [run-name]. Use the nextflow log command to show previous run names.

-c

Specify the path to a specific config file (this is a core Nextflow command). See the nf-core website documentation for more information.

Custom configuration

Resource requests

Whilst the default requirements set within the pipeline will hopefully work for most people and with most input data, you may find that you want to customise the compute resources that the pipeline requests. Each step in the pipeline has a default set of requirements for number of CPUs, memory and time. For most of the pipeline steps, if the job exits with any of the error codes specified here it will automatically be resubmitted with higher resources request (2 x original, then 3 x original). If it still fails after the third attempt then the pipeline execution is stopped.

To change the resource requests, please see the max resources and customise process resources section of the nf-core website.

Custom Containers

In some cases, you may wish to change the container or conda environment used by a pipeline steps for a particular tool. By default, nf-core pipelines use containers and software from the biocontainers or bioconda projects. However, in some cases the pipeline specified version maybe out of date.

To use a different container from the default container or conda environment specified in a pipeline, please see the updating tool versions section of the nf-core website.

Custom Tool Arguments

A pipeline might not always support every possible argument or option of a particular tool used in pipeline. Fortunately, nf-core pipelines provide some freedom to users to insert additional parameters that the pipeline does not include by default.

To learn how to provide additional arguments to a particular tool of the pipeline, please see the customising tool arguments section of the nf-core website.

nf-core/configs

In most cases, you will only need to create a custom config as a one-off but if you and others within your organisation are likely to be running nf-core pipelines regularly and need to use the same settings regularly it may be a good idea to request that your custom config file is uploaded to the nf-core/configs git repository. Before you do this please can you test that the config file works with your pipeline of choice using the -c parameter. You can then create a pull request to the nf-core/configs repository with the addition of your config file, associated documentation file (see examples in nf-core/configs/docs), and amending nfcore_custom.config to include your custom profile.

See the main Nextflow documentation for more information about creating your own configuration files.

If you have any questions or issues please send us a message on Slack on the #configs channel.

Running in the background

Nextflow handles job submissions and supervises the running jobs. The Nextflow process must run until the pipeline is finished.

The Nextflow -bg flag launches Nextflow in the background, detached from your terminal so that the workflow does not stop if you log out of your session. The logs are saved to a file.

Alternatively, you can use screen / tmux or similar tool to create a detached session which you can log back into at a later time. Some HPC setups also allow you to run nextflow within a cluster job submitted your job scheduler (from where it submits more jobs).

Nextflow memory requirements

In some cases, the Nextflow Java virtual machines can start to request a large amount of memory. We recommend adding the following line to your environment to limit this (typically in ~/.bashrc or ~./bash_profile):

NXF_OPTS='-Xms1g -Xmx4g'

Troubleshooting

If you experience any issues, please open an issue on our GitHub repository. However, some resolutions for common issues will be noted below:

  • Due to the nature of the data this pipeline analyzes, some tools may experience increased runtimes. For some of the custom tools made for this pipeline (preextract_fastq.py and correct_barcodes.py), we have leveraged the splitting done via the split_amount parameter to decrease their overall runtimes. The split_amount parameter will split the input FASTQs into a number of FASTQ files, each containing a number of lines based on the value used for this parameter. As a result, it is important not to set this parameter to be too low as doing so would cause the creation of a large number of files the pipeline will be processed. While this value can be highly dependent on the data, a good starting point for an analysis would be to set this value to 500000. If you find that PREEXTRACT_FASTQ and CORRECT_BARCODES are still taking long amounts of time to run, it would be worth reducing this parameter to 200000 or 100000, but keeping the value on the order of hundred of thousands or tens of thousands should help with keeping the total number of processes minimal. An example of setting this parameter to be equal to 500000 is shown below:
split_amount: 500000
  • We have seen a recurrent node failure on slurm clusters that does seem to be related to submission of Nextflow jobs. This issue is not related to this pipeline per se, but rather to Nextflow itself. We are currently working on a resolution. But we have two methods that appear to help overcome should this issue arise:
    1. Provide a custom config that increases the memory request for the job that failed. This may take a couple attempts to find the correct requests, but we have noted that there does appear to be a memory issue occasionally with these errors.
    2. Request an interactive session with a decent amount of time and memory and CPUs in order to run the pipeline on the single node. Note that this will take time as there will be minimal parallelization, but this does seem to resolve the issue.
  • By default the pipeline groups UMIs by feature rather than by alignment position: --dedup_per_gene (genome alignments) and --dedup_per_contig (transcriptome alignments), both true. This matters because umi_tools corrects its bundling position for soft clipping, and long reads carry variable amounts of untrimmed adapter and polyA, so reads from one PCR family end up in different position groups even when they align at an identical coordinate. Position-based grouping consequently leaves a large fraction of duplicates in place, and absolute molecule counts come out inflated.
    • --dedup_per_gene tags every alignment with the gene whose body it overlaps most (GX/GN/GS tags, written by TAG_GENES) and groups on that gene. It requires --gtf. Reads whose gene call is ambiguous, or that overlap no gene at all, are deduplicated by position instead and merged back in, so no reads are dropped. The GS distribution is published under qc/gene_assignment/.
    • --dedup_per_contig applies to transcriptome alignments only, where a contig is a transcript. It needs no gene tag and no fallback.
    • Set either to false to restore position-based grouping. Note that these change absolute UMI counts substantially. Relative expression is essentially unaffected, but anything quoting UMIs per cell, library complexity or saturation will differ from earlier releases.
  • We note that umitools dedup can take a large amount of time in order to perform deduplication. One approach we have implemented to assist with speed is to split input files based on chromosome. However for the transcriptome aligned bams, there is some additional work required that involves grouping transcripts into appropriate chromosomes. In order to accomplish this, the pipeline needs to parse the transcript id from the transcriptome FASTA file. The transcript id is often nested in the sequence identifier with additional data and the data is delimited. We have included the delimiters used by reference files obtained from GENCODE, NCBI, and Ensembl. However in case you wish to explicitly control this or if the reference file source uses a different delimiter, you are able to manually set it via the --fasta_delimiter parameter.
  • We acknowledge that analyzing PromethION data is a common use case for this pipeline. Currently, the pipeline has been developed with defaults to analyze GridION and average sized PromethION data. For cases, where jobs have fail due for larger PromethION datasets, the defaults can be overwritten by a custom configuation file (provided by the -c Nextflow option) where resources can be increased (substantially in some cases). Below are some of the overrides we have used, and while these amounts may not work on every dataset, these will hopefully at least note which processes will need to have their resources increased:
process
{
    withName: '.*:.*FASTQC.*'
    {
        cpus = 20
    }
}

process
{
    withName: '.*:BLAZE'
    {
        cpus = 30
    }
}

process
{
    withName: '.*:TAG_BARCODES'
    {
        memory = '60.GB'
    }
}

process
{
    withName: '.*:SAMTOOLS_SORT'
    {
        cpus = 20
    }
}

process
{
    withName: '.*:MINIMAP2_ALIGN'
    {
        cpus = 20
    }
}

process
{
    withName: '.*:ISOQUANT'
    {
        cpus = 30
        memory = '85.GB'
    }
}

We further note that while we encourage the use of split_amount as discussed above for larger datasets, the pipeline can be executed without enabling this parameter. When doing this, please consider increasing the time limit to CORRECT_BARCODES as it can take hours instead of minutes when split_amount is disabled:

//NOTE: with split_amount disabled, consider increasing the time limit to CORRECT_BARCODES
process
{
    withName: '.*:CORRECT_BARCODES'
    {
        time = '15.h'
    }
}