search

friend1ws / nanomonsv

SV detection tool for nanopore sequence reads
GNU General Public License v3.0
88 stars 12 forks source link

readme

nanomonsv

License: GPL v3 Build Status

Introduction

nanomonsv is a software for detecting somatic structural variations from paired (tumor and matched control) cancer genome sequence data. nanomonsv is presented in the following paper. When you use nanomonsv or any resource of this repository, please kindly site this paper.

Precise characterization of somatic complex structural variations from tumor/control paired long-read sequencing data with nanomonsv, Shiraishi et al., Nucleic Acids Research, 2023, [link].

The current version of nanomonsv includes two detection modules, Canonical SV module, and Single breakend SV module. Canonical SV module can identify somatic SVs that can be captured by short-read technologies with higher precision and recall than existing methods. Furthermore, Single breakend SV module enables the detection of complex SVs that can only be identified by long-reads, such as SVs involving highly-repetitive centromeric sequences, and LINE1- and virus-mediated rearrangements.

Please see the wiki page for Single breakend SV module.

Dependency

For basic use (parse, get command)

Binary programs

htslib, mafft, racon(optional from ver. 0.3.0. However, we recommend to use this option. Add --use_racon option when you perfrom get command.)

Python

Python (tested with >=3.6), pysam, numpy, parasail

For advanced use (insert_classify command)

bwa, minimap2, bedtools, RepeatMasker

Preparation

For basic use (parse, get command)

Install software and add them to the PATH

nanomonsv uses, tabix, bgzip (which are part of HTSlib projects) and mafft inside the program, assuming those are installed, and the paths are already added to the running environment.

For use of racon

Since version 0.3.0, we support racon for the step where generating consensus sequence and get single-base resolution breakpoints. racon may become the default instead of mafft in the future.

For advanced use (insert_classify command)

bwa, minimap2, bedtools and RepeatMasker are required to be installed and these paths are added to the running environment.

Input file

nanomonsv accept the BAM file aligned by minimap2.

Control panel

Starting with version 0.5.0, the use of the control panel is supported. In this software, supporting reads for SVs are collected for multiple samples other than the target sample, and such reads are removed as common noise (or those derived from common SVs) in the get stage. This strategy is expected to exclude many false positives as well as improve computational cost.

We have prepared the command to create control panels from the user's own sequencing data. In addition, for users who do not have sufficient sequencing data that can serve as a control panel (or just do not have time for processing), we prepared a control panel that has been created using the 30 Nanopore sequencing data from the Human Pangenome Reference Consortium, which is available at zenodo.

This control panel is made by aligning 30~40 Nanopore (basecalled by Guppy ver. 4 and 6) and PacBio HiFi sequencing data to the GRCh38/CHM13 reference genome (obtained from here) with minimap2 version 2.24. For the choice of the control panel, we recommend that you use one that is as close as possible in platform and basecall quality. However, based on our experience, a noisier control panel tends to be more versatile. Therefore, when unsure, we advise to use the Nanopore control panel from Guppy version 4. When you use these control panels and publish, do not forget to credit to HPRC!

Quickstart

  1. Install all the prerequisite software and install nanomonsv. 
    pip install nanomonsv (--user)

You can also install nanomonsv via conda (bioconda channel).

conda create -n nanomonsv -c conda-forge -c bioconda nanomonsv

Occasionally the conda releases lag behind the source code and PyPI releases.

  1. Prepare the reference genome for the test data (here, we show the path to Genomic Data Commons reference genome).

    wget https://api.gdc.cancer.gov/data/254f697d-310d-4d7d-a27b-27fbf767a834 -O GRCh38.d1.vd1.fa.tar.gz
tar xvf GRCh38.d1.vd1.fa.tar.gz

  2. Parse the putative structural variation supporting reads of the test data.

    nanomonsv parse tests/resource/bam/test_tumor.bam output/test_tumor
nanomonsv parse tests/resource/bam/test_ctrl.bam output/test_ctrl

  3. Get the final result.

    nanomonsv get output/test_tumor tests/resource/bam/test_tumor.bam GRCh38.d1.vd1.fa --control_prefix output/test_ctrl --control_bam tests/resource/bam/test_ctrl.bam

You will see the result file named test_tumor.nanomonsv.result.txt.

Realistic example sequencing data

The Oxford Nanopore Sequencing data used in the bioRxiv paper is available through the public sequence repository service (BioProject ID: PRJDB10898):

The results of nanomonsv for the above data are available here. When you perform nanomonsv to the above data and have experienced errors, please report to us. Also, please kindly cite the NAR paper when you use these data.

See tutorial wiki page for an example workflow on analyzing COLO829 sample.

Commands

parse

This step parses all the supporting reads of putative somatic SVs.

nanomonsv parse [-h] [--reference_fasta reference.fa] [--debug]
                [--split_alignment_check_margin SPLIT_ALIGNMENT_CHECK_MARGIN]
                [--minimum_breakpoint_ambiguity MINIMUM_BREAKPOINT_AMBIGUITY]
                alignment_file output_prefix

See the help (nanomonsv parse -h) for other options. From v0.7.0, nanomonsv can accept CRAM format files. For CRAM files, we recommend to add the PATH to the reference genome file by --reference_fasta.

After successful completion, you will find supporting reads stratified by deletions, insertions, and rearrangements: ({output_prefix}.deletion.sorted.bed.gz, {output_prefix}.insertion.sorted.bed.gz, {output_prefix}.rearrangement.sorted.bedpe.gz, and {output_prefix}.bp_info.sorted.bed.gz and their indexes (.tbi files).

get

This step gets the SV result from the parsed supporting reads data obtained above.

nanomonsv get [-h] [--control_prefix CONTROL_PREFIX]
              [--control_bam CONTROL_BAM]
              [--min_tumor_variant_read_num MIN_TUMOR_VARIANT_READ_NUM]
              [--min_tumor_VAF MIN_TUMOR_VAF]
              [--max_control_variant_read_num MAX_CONTROL_VARIANT_READ_NUM]
              [--max_control_VAF MAX_CONTROL_VAF]
              [--cluster_margin_size CLUSTER_MARGIN_SIZE]
              [--median_mapQ_thres MEDIAN_MAPQ_THRES]
              [--max_overhang_size_thres MAX_OVERHANG_SIZE_THRES]
              [--var_read_min_mapq VAR_READ_MIN_MAPQ]
              [--qv10] [--qv15] [--qv20] [--qv25] [--use_racon]
              [--single_bnd] [--threads THREADS] [--processes PROCESSES] 
              [--sort_option SORT_OPTION] [--max_memory_minimap2] [--debug]
              tumor_prefix tumor_bam reference.fa

This software can generate a list of SVs without specifying the matched control. But we have not tested the performance of the approach just using tumor sample, and strongly recommend using the matched control data. Even when only tumor sample is available, we still recommend using dummy control sample (collected from other person's tissue).

When you use the control panel (recommended!), use the following argument.

You can also use --process to use multi-processing mode. Currently, we do not recommend using --thread option.

From v0.7.0, we prepared preset parameter options:

After successful execution, you will be able to find the result file names as {tumor_prefix}.nanomonsv.result.txt. See the help (nanomonsv get -h) for other options.

When you want to change the engine of Smith-Waterman algorithm to SSW Library, specify --use_ssw_lib option, though we do not generally recommend this.

Also, we basically recommend using --use_racon option. This will slightly improve the identification of single-base resolution breakpoint, and polishing of inserted sequences.

For detection of single breakend SVs, please use --single_bnd option as well as --use_racon. Please see wiki page.

Also, we have prepared the script (misc/post_fileter.py) for filtering the result. Please see the wiki page. For output files of the version 0.4.0 and later, some filtering has already been performed (see the wiki page). However, we strongly recommed to perform additional processing; removing indels within simple repeat regions (see the wiki page).

From the version 0.4.0, we will also provide the VCF format result files.

result

merge_control

This command merges non-matched control panel supporting reads obtained by performing parse command.

nanomonsv merge_control [-h] prefix_list_file output_prefix

insert_classify

This command classifies the long insertions into several mobile element insertions (still in alpha version). This does not yet support VCF format, but we will do so in the near future.

nanomonsv insert_classify [-h] [--grc] [--genome_id {hg19,hg38,mm10}]
                          [--debug]
                          sv_list_file output_file reference.fa

result

validate

This command, which is part of the procedures of get command, performs validation of the candidate SVs by alignment of tumor and matched control BAM files. This may be helpful for the evaluation of SV tools of the short-read platform when pairs of short-read and long-read sequencing data are available. This is still in alpha version.

nanomonsv validate [-h] [--control_bam CONTROL_BAM]
                   [--var_read_min_mapq VAR_READ_MIN_MAPQ] [--debug]
                   sv_list_file tumor_bam output reference.fa