Skip to content

Latest commit

 

History

History
172 lines (120 loc) · 7.9 KB

README.md

File metadata and controls

172 lines (120 loc) · 7.9 KB

Whisper2

GitHub downloads GitHub Actions CI License: GPL v3

Quick start

# download and unpack E.coli str. K-12 substr. MG1655 reference genome
wget https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/904/425/475/GCF_904425475.1_MG1655/GCF_904425475.1_MG1655_genomic.fna.gz
gzip -k -d GCF_904425475.1_MG1655_genomic.fna.gz 

# download NovaSeq 6000 reads from E.coli MG1655 IR-10-94 population sequencing
wget ftp://ftp.sra.ebi.ac.uk/vol1/fastq/SRR100/030/SRR10051130/SRR10051130_1.fastq.gz
wget ftp://ftp.sra.ebi.ac.uk/vol1/fastq/SRR100/030/SRR10051130/SRR10051130_2.fastq.gz

# clone and build Whisper
git clone https://github.com/refresh-bio/Whisper
make -C Whisper/src

# build an index named ecoli for the reference genome
mkdir index
mkdir temp
./Whisper/src/whisper-index ecoli GCF_904425475.1_MG1655_genomic.fna ./index ./temp

# map reads to the reference genome and store the results in mappings.sam
./Whisper/src/whisper -rp -out mappings -temp ./temp/ ./index/ecoli  SRR10051130_1.fastq.gz SRR10051130_2.fastq.gz

Note, that Whisper was optimized for processing data from sequencing large samples with high coverage (e.g. human).

Installation and configuration

Whisper comes with a set of precompiled binaries for Windows and Linux. They can be found under Releases tab.

The software can be also built from the sources distributed as:

  • Visual Studio 2017 solution for Windows,
  • MAKE project (G++ 7.2 required) for Linux.

NOTE

Linux systems limit number of files that can be opened by the process. Make sure this limit is sufficient for Whisper, which requires:

num_files = num_bins (384 by default) + num_threads + 2 * total_size_of_FASTQ_in_GB

E.g., if sample reads in uncompressed FASTQ format have 100GB and the processing is done by 16 CPU threads, Whisper opens 600 files. To change the limit use ulimit Linux command:

ulimit -n 600

Usage

The preliminary step of the analysis, performed only once for a given reference genome, is construction of an index. The index may be then used for mapping reads from different samples to the reference.

Indexing reference genome

Indexing can be executed in two wariants, depending on the representation of the reference (single versus multiple FASTA files):

whisper-index <index_name> <ref_seq_file_name> <dest_dir> <temp_dir>

whisper-index <index_name> @<ref_seq_files_name> <dest_dir> <temp_dir>

Parameters:

  • index_name - name of index (output),
  • ref_seq_file_name - reference sequence in the FASTA format (input)
  • ref_seq_files_name - file with a list of multiple reference FASTA files (input)
  • dest_dir - existing destination directory where index files should be placed (output)
  • temp_dir- existing temporary directory (output)

Examples:

  • whisper-index hg38-chr20 chr20.fa index-dir temp-dir

Generates index named hg38-chr20 for chr20.fa reference sequence and places it in index-dir directory.

  • whisper-index hg38 @hg38.list index-dir temp-dir

Generates index named hg38 for all FASTA files listed in hg38.list file and places it in index-dir directory.

Mapping reads

whisper [options] <index_name> @<files>

whisper [options] <index_name> file_se

whisper [options] <index_name> file_pe_1 file_pe_2

Parameters:

  • index_name - name of the index (as created by whisper-index)
  • files - name of the file containing list of FASTQ files with seq. reads
  • file_se - FASTQ file (single-end)
  • file_pe_[1|2] - FASTQ files (paired-end)

Options:

  • -b <value> - no. of temporary files (minimum: 100, default: 384)
  • -clipping-distance <value> - no. of sigmas for max. additional distance in clipping (default: 14)
  • -d[fr/ff/rf] - mapping orientation (default: -dfr (forward - reverse)
  • -dist_paired <value> - max. distance for paired read (default: 1000)
  • -e <value> - max. no of errors (default: auto)
  • -e-paired <value> - max. fraction of errors in paired read (default: 0.09)
  • -enable-boundary-clipping <value> - enable clipping at boundaries when a lot of mismatches appears (default: 1)
  • -enable-mapping_indels <value> - enable looking for long indels during mapping stages (default: 1)
  • -enable-short-indel-refinement <value> - enable short indel refinement after mapping (default: 1)
  • -enable-short-reads <value> - enable reads shorter than 90% of the longest reads (default: 0)
  • -filter <value> - store only mappings for given chromosome (default: )
  • -gap-del-open <value> - score for gap (del) open (default: -5)
  • -gap-del-extend <value> - score for gap (del) extend (default: -0.4)
  • -gap-ins-open <value> - score for gap (ins) open (default: -5)
  • -gap-ins-extend <value> - score for gap (ins) extend (default: -0.4)
  • -gzipped-SAM <value> - gzip compression level of SAM/BAM, 0 - no compression (default: 0)
  • -high-confidence-sigmas <value> - (default: 4)
  • -hit-merging-threshold <value> - minimal distance between different mappings (default: 12)
  • -hit-merging-wrt-first <value> - calculate distance in marged group w.r.t. first (default: 1)
  • -m[f/s/a] - mode: first stratum/second stratum/all strata (default: first stratum)
  • -mask-lqb <value> - mask bases of quality lower than value (default: 0)
  • -max-indel-len <value> - max. indel length (default: 50)
  • -min-clipped-factor <value> - mask bases of quality lower than value (default: 1)
  • -out <name> - name of the output file (default: whisper)
  • -penalty-saturation <value> - no. of sigmas for max. penalty in matching pairs (default: 7)
  • -rg <read_group> - complete read group header line, '\t' character will be converted to a TAB in the output SAM while the read group ID will be attached to every read (example line: '@RG ID:foo SM:bar')
  • -r[s|p] - single or paired-end reads (default: single)
  • -score-discretization-threshold (default: 0.5)
  • -score-clipping <value> score for clipping (default: -6)
  • -score-match <value> - score for matching symbol (default: 1)
  • -score-mismatch <value> - score for mismatching symbol (default: -5)
  • -sens <value> - turn on/off sensitive mode (default: 1)
  • -sens-factor <value> - sensitivity factor (default: 2.5)
  • -stdout - use stdout to store the output (default: 0)
  • -store-BAM - turn on saving in BAM (default: 0)
  • -t <value> - no. of threads (0-adjust to hardware) (default: 0)
  • -temp <name> - prefix for temporary files (default: ./whisper_temp_)
  • -x - load complete suffix arrays in main memory (default: 0)

Examples:

  • whisper -out result.sam -rp -t 12 hg38 reads_1.fq reads_2.fq

Maps paired-end reads from reads_1.fq and reads_2.fq FASTQ files using hg38 index. Computations are distributed over 12 threads, results are stored in result.sam file.

  • whisper hg38 @reads_se.list

Maps single-end reads from FASTQ files listed in reads_se.list file using hg38 index. The example contents of reads_se.list file:

readsA
readsB
readsC
...
  • whisper -rp hg38 @reads_pe.list

Maps paired-end reads from FASTQ files listed in reads_pe.list file using hg38 index. The example contents of reads_pe.list file:

readsA_1;readsA_2
readsB_1;readsB_2
readsC_1;readsC_2
...

Citing

Deorowicz, S., Debudaj-Grabysz, A., Gudyś, A., Grabowski, S. (2018) Whisper: Read sorting allows robust mapping of sequencing data, Bioinformatics, 35(12):2043–2050

Deorowicz, S., Gudyś, A. (2021) Whisper 2: Indel-sensitive short read mapping, SoftwareX, 14:100692