nil-nf pipeline will align, call variants, and generate datasets for NIL sequence data. It runs a hidden-markov-model to fill in missing genotypes from low-coverage sequence data.
███╗ ██╗██╗██╗ ███╗ ██╗███████╗ ████╗ ██║██║██║ ████╗ ██║██╔════╝ ██╔██╗ ██║██║██║█████╗██╔██╗ ██║█████╗ ██║╚██╗██║██║██║╚════╝██║╚██╗██║██╔══╝ ██║ ╚████║██║███████╗ ██║ ╚████║██║ ╚═╝ ╚═══╝╚═╝╚══════╝ ╚═╝ ╚═══╝╚═╝ parameters description Set/Default ========== =========== ======= --debug Set to 'true' to test false --cores Number of cores 4 --A Parent A N2 --B Parent B CB4856 --cA Parent A color (for plots) #0080FF --cB Parent B color (for plots) #FF8000 --out Directory to output results NIL-N2-CB4856-2017-09-27 --fqs fastq file (see help) (required) --reference Reference Genome /Users/dancook/Documents/git/nil-nf/reference/WS245.fa.gz --vcf VCF to fetch parents from (required) --tmpdir A temporary directory tmp/ The Set/Default column shows what the value is currently set to or would be set to if it is not specified (it's default).
The pipeline comes pre-packed with fastq's and a VCF that can be used to debug. You can use the following command to debug:
nextflow run main.nf --debug --reference=<path to reference>
The number of cores to use during alignments and variant calling.
Two parental strains must be provided. By default these are N2 and CB4856. The parental strains provided must be present in the VCF provided. Their genotypes are pulled from that VCF and used to generate the HMM. See below for more details.
The color to use for parental strain A and B on plots.
A directory in which to output results. By default it will be
NIL-A-B-YYYY-MM-DD where A and be are the parental strains.
In order to process NIL data, you need to move the sequence data to a folder and create a
fq_sheet.tsv. This file defines the fastqs that should be processed. The fastq files are relative to that file. The fastq sheet details the FASTQ files and their associated strains. It should be tab-delimited and look like this:
NIL_01 NIL_01_ID S16 NIL_01_1.fq.gz NIL_01_2.fq.gz NIL_02 NIL_02_ID S1 NIL_02_1.fq.gz NIL_02_2.fq.gz
Notice that the file does not include a header. The table with corresponding columns looks like this.
The columns are detailed below:
- strain - The name of the strain. If a strain was sequenced multiple times this file is used to identify that fact and merge those fastq-pairs together following alignment.
- fastq_pair_id - This must be unique identifier for all individual FASTQ pairs.
- library - A string identifying the DNA library. If you sequenced a strain from different library preps it can be beneficial when calling variants. The string can be arbitrary (e.g. LIB1) as well if only one library prep was used.
- fastq-1-path - The relative path of the first fastq.
- fastq-2-path - The relative path of the second fastq.
This file needs to be placed along with the sequence data into a folder. The tree will look like this:
NIL_SEQ_DATA/ ├── NIL_01_1.fq.gz ├── NIL_01_2.fq.gz ├── NIL_02_1.fq.gz ├── NIL_02_2.fq.gz └── fq_sheet.tsv
Do not perform any pre-processing on NIL data. NIL-data is low-coverage by design and you want to retain as much sequence data (however poor) as possible.
--vcf (Parental VCF)¶
Before you begin, you will need access to a VCF with high-coverage data from the parental strains. In general, this can be obtained using the latest release of the wild-isolate data which is usually located in the b1059 analysis folder. For example, you would likely want to use:
This is the hard-filtered VCF, meaning that poor quality variants have been stripped. Use hard-filtered VCFs for this pipeline.
Set the parental VCF as
A fasta reference indexed with BWA. On Quest, the reference is available here:
A directory for storing temporary data.
The final output directory looks like this:
. ├── log.txt ├── fq │ ├── fq_bam_idxstats.tsv │ ├── fq_bam_stats.tsv │ ├── fq_coverage.full.tsv │ └── fq_coverage.tsv ├── SM │ ├── SM_bam_idxstats.tsv │ ├── SM_bam_stats.tsv │ ├── SM_coverage.full.tsv │ └── SM_coverage.tsv ├── hmm │ ├── gt_hmm.(png/svg) │ └── gt_hmm.tsv ├── bam │ └── <BAMS + indices> ├── duplicates │ └── bam_duplicates.tsv ├── sitelist │ ├── N2.CB4856.sitelist.tsv.gz │ └── N2.CB4856.sitelist.tsv.gz.tbi └── vcf ├── NIL.filtered.stats.txt ├── NIL.filtered.vcf.gz ├── NIL.filtered.vcf.gz.csi ├── NIL.hmm.vcf.gz ├── NIL.hmm.vcf.gz.csi ├── gt_hmm.tsv ├── gt_hmm_fill.tsv └── union_vcfs.txt
A summary of the nextflow run.
bam_duplicates.tsv - A summary of duplicate reads from aligned bams.
- fq_bam_idxstats.tsv - A summary of mapped and unmapped reads by fastq pair.
- fq_bam_stats.tsv - BAM summary by fastq pair.
- fq_coverage.full.tsv - Coverage summary by chromosome
- fq_coverage.tsv - Simple coverage file by fastq
If you have multiple fastq pairs per sample, their alignments will be combined into a strain or sample-level BAM and the results will be output to this directory.
- SM_bam_idxstats.tsv - A summary of mapped and unmapped reads by sample.
- SM_bam_stats.tsv - BAM summary at the sample level
- SM_coverage.full.tsv - Coverage at the sample level
- SM_coverage.tsv - Simple coverage at the sample level.
- gt_hmm.(png/svg) - Haplotype plot for NILs.
- gt_hmm.tsv - Long form genotypes file.
- coverage_comparison.(png/svg/pdf) - Compares FASTQ and Sample-level coverage. Note that coverage is not simply cumulative. Only uniquely mapped reads count towards coverage, so it is possible that the sample-level coverage will not equal to the cumulative sum of the coverages of individual FASTQ pairs.
- duplicates.(png/svg/pdf) - Coverage vs. percent duplicated.
- unmapped_reads.(png/svg/pdf) - Coverage vs. unmapped read percent.
<A>.<B>.sitelist.tsv.gz[+.tbi]- A tabix-indexed list of sites found to be different between both parental strains.
- gt_hmm.tsv - Haplotypes defined by region with associated information.
- gt_hmm_fill.tsv - Same as above, but using
--endfillwith VCF-Kit. For more information, see VCF-Kit Documentation
- NIL.filtered.vcf.gz - A VCF genotypes including the NILs and parental genotypes.
- NIL.filtered.stats.txt - Summary of filtered genotypes. Generated by
bcftools stats NIL.filtered.vcf.gz
- NIL.hmm.vcf.gz - The RIL VCF as output by VCF-Kit; HMM applied to determine genotypes.
- union_vcfs.txt - A list of VCFs that were merged to generate RIL.filter.vcf.gz