repository·develop·Indexed 24 days ago
https://github.com/samtools/samtoolsA suite of tools for manipulating and analyzing high-throughput sequencing data in SAM, BAM, and CRAM formats. SAMtools works in conjunction with htslib and bcftools. The toolkit includes specialized commands such as addreplacerg for managing read group tags, ampliconclip for clipping read alignments, and ampliconstats for producing amplicon sequencing statistics.
SAMtools is part of a coordinated suite of tools for handling high-throughput sequencing data. It is one of three primary projects:
mpileup).Because release tarballs contain generated files not present in the Git repository, you must perform extra steps to build from source.
By default, the build process looks for an HTSlib source tree in ../htslib. To use a different HTSlib source tree or a previously installed version, use the --with-htslib=DIR flag during the configuration step.
autoheader # Build config.h.in
autoconf -Wno-syntax # Generate the configure script
./configure # Configure optional functionality (use --with-htslib=DIR to specify HTSlib path)
make
make installSamtools is a suite of utilities designed to manipulate sequence alignments in SAM (Sequence Alignment/Map), BAM, and CRAM formats. It supports format conversion, sorting, merging, indexing, and rapid retrieval of reads from specific genomic regions.
Key characteristics:
- represents stdin, and an output file of - represents stdout. This allows for efficient use of Unix pipes.file.bam.bai), it tries without the suffix (e.g., file.bai). You can explicitly link a data file and an index in a different location using the ##idx## notation: /path/to/data.bam##idx##/path/to/index.bai.The samtools samples command extracts sample names from read-group (@RG) headers and identifies the associated reference genome paths from SAM, BAM, or CRAM files. The output is a tab-delimited list suitable for workflow managers.
Output Format:
.) is used..) is used. This column is populated if you provide references via -f or -F.Y/N) indicating if the file is indexed. This column is populated if you use the -i flag.When converting paired-end data, samtools fastq categorizes reads based on the READ1 and READ2 flag bits:
Managing Output Files:
-1 FILE for READ1, -2 FILE for READ2, or -o FILE (equivalent to -1 and -2) to write specific categories to files.-s FILE to write singleton reads (reads that do not have a corresponding pair in the other category) to a specific file. If -s is used, only paired sequences will be output to the standard output/other files for categories 1 and 2.samtools fastq ... > all_reads.fq).samtools fastq -0 /dev/null -s single.fq -N in_name.bam > paired.fqThe samtools rmdup command removes potential PCR duplicates. If multiple read pairs have identical external coordinates, it retains only the pair with the highest mapping quality.
Important Constraints:
ISIZE to be correctly set. It does not work for unpaired reads (e.g., orphan reads or ends mapped to different chromosomes).-s flag.If you need to handle unpaired reads (orphan reads or ends on different chromosomes), use Picard's MarkDuplicates instead, as samtools rmdup is limited in this regard.
When using samtools view to output CRAM, you can use the --output-fmt-option to set a compression profile. This simplifies setting multiple encoding options at once.
Profiles available:
fast: Optimized for speed (low compression).normal: The default profile.small: Optimized for smaller file size.archive: Optimized for maximum compression (uses more intensive methods like use_arith).Example:
samtools view -O cram,small -o bar.cram bar.bamsamtools view -O cram,small -o bar.cram bar.bamFast mode (-f) is an optimized version of collate that uses an in-memory buffer to write out most pairs as soon as they are found, avoiding excessive temporary file usage.
Key Characteristics of Fast Mode:
READ1 or READ2 flags set (but not both). Any other alignment records are filtered out. This requires that there are no more than two reads for any given QNAME after filtering.Tuning Fast Mode:
-r INT to specify the number of reads to store in memory. Increasing this value uses more memory but allows more pairs to be written out early.samtools idxstats can run on SAM files, CRAM files, or unindexed BAM files, it will be significantly slower in those cases. When the input is unindexed, the command must read through the entire file to produce the summary statistics instead of using the BAM indices.The -P, --fetch-pairs option retrieves pairs even when the mate is outside the requested region.
Requirements and Behavior:
-M).-L).RNEXT/PNEXT fields, then a second pass to collect the actual reads.-c, -U, or -p.RNEXT and PNEXT fields must be accurate (use samtools fixmate if necessary).The --split TAG option allows you to generate categorized statistics based on a specific SAM tag (e.g., RG for Read Group).
When using this option, samtools stats will create additional files named <prefix>_<value>.bamstat for every unique value encountered in that tag. You can use --split-prefix STR to define a custom prefix for these files. If no prefix is provided, the input filename is used.
The output contains file-specific sections (prefixed with F) and combined sections (prefixed with C). File-specific sections are interleaved and ordered by file, then by statistic type. To extract a specific statistic across all files, use grep.
Key Output Sections:
-m).VDEPTH variants only include "valid" templates (full-length coverage).-d).VALID indicates templates with full-length coverage and matching primers. Uses run-length encoding (controlled by -D).