vg 1.60.0 - Annicco

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.60.0

Buildable Source Tarball: vg-v1.60.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• vg's self-reported version information should no longer be able to lag behind the actual Git revision checked out
• vg stats --snarl-contents added to print a table of snarl / depth / parent / nested node ids
• vg inject can now handle reads that softclip after the end of the target path
• vg paths -n option added to normalize graphs using path information to "snap together" redundant paths through snarls. After running, no two path traversals through a snarl will ever produce the same sequence string without the traversals themselves being identical.
• vg build system no longer patches kff-cpp-api CMakeLists.txt.

Updated Submodules

The kff-cpp-api, libbdsg and sublinear-Li-Stephens submodules have been updated.

vg 1.59.0 - Casatico

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.59.0

Buildable Source Tarball: vg-v1.59.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• Experimental haploid scoring model for haplotype sampling.
• Faster BAM output when using a GBZ format graph
• vg inject no longer reports invalid alignments
• vg giraffe --parameter-preset long-form option should now work again
• vg surject now complains if extra command line arguemnts are provided
• vg Mac builds no longer pass -headerpad without an argument to the linker
• vg giraffe in non-chaining single-end mode will again prioritize minimizers the same way as in paired-end mode
• README no longer interpretable as suggesting that users build with very old Clang
• vg surject now uses an unlimited --max-anchors by default
• vg clip faster on graphs with high path depth
• vg surject has improved base-level alignment accuracy with long reads
• Random walk band padding is now available outside MultipathAligner
• Gapless extension queue order should no longer cause vg giraffe output to very across compilers
• vg giraffe --show-work no longer claims to not be doing rescue when it is
• vg annotate can now output alignments in GAF.
• When reading more than 1000 BED or GFF records, vg will no longer dump the first records to standard output and forget about them.

Updated Submodules

The libbdsg and libhandlegraph submodules have been updated.

vg 1.58.0 - Cartari

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.58.0

Buildable Source Tarball: vg-v1.58.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• vg deconstruct now does path-based (formerly -e) deconstruction by default. Old default behaviour of exhaustively processing (tiny) sites is deprecated.
• if -a is not used, vg deconstruct will recursive on child snarls of snarls it failed to process (like vg call) - functionality that was, I think, dropped a while back.
• Experimental option -L added to vg deconstruct in order to cluster similar allele traversals together. The value given is a (length-weighted) threshold for the jaccard coefficient between the oriented nodes of two traversals. So if -L 0.75 is given, then alleles that have >= 0.75 similarity based on their graph positions will be merged into one. Two new FORMAT fields are added to keep track of the difference, TS (jaccard distance) and TL (length difference). Clustering is done greedily starting with selected reference paths.
• new (experimental) option -n added to vg deconstruct. Like -a, it genotypes nested sites, but unlike -a it does so top-down, setting various tags that keep track of the nesting relationship at the allele level (and also linking every site back to its position on the LV=0 reference chromosome). *-alleles (used in recent VCF versions to represent spanning alleles) are used. This option will not support nested insertions on GBZ/GBWT input -- so in practice it should be used on chromosome-level .vg files (I will look into relaxing this).
• -R option added to vg deconstruct to toggle whether star-alleles are reported with -n.
• README now explains how to get vg on your PATH
• README now explains how to build on multiple threads
• vg can now read GAM files generated by the long-read Giraffe prototype
• vg filter now lets you require exact matches for name filters instead of prefix matches with --exact-name.
• deconstruct/call can write giant VCF lines. This happens in, say, large svs with lots of samples that each get their own allele due to nested variation (hopefully deconstruct -L can mitigate this via merging). Giant AT fields for each allele don't help. bcf apparently has a 2 gig line limit, and there's a case of deconstruct seemingly truncating large records. vg deconstruct / call are now modified to drop (with a warning) any lines >2Gb to avoid these issues.
• vg giraffe should no longer crash when mapping paired-end reads and reporting secondaries without a fragment length distribution
• vg inject now supports GAF format with the new --output-format/-o option

Updated Submodules

• libvgio

vg 1.57.0 - Franchini

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.57.0

Buildable Source Tarball: vg-v1.57.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• Added haplotype sampling example
• Added subpath support to vg annotate
• Duplicate words removed from documentation
• WFAExtender alignments are now scored using the full length bonus
• Extra debug messages removed from vg chunk
• libdeflate build should now have more space to be renamed

Updated Submodules

None

vg 1.56.0 - Collalto

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.56.0

Buildable Source Tarball: vg-v1.56.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• vg GFA output should now use P lines instead of W lines when nonzero phase blocks are present, to allow them to be round-tripped.
• vg call -c and -C options changed to limit search based on all alleles and not just reference allele. This means these options work much better in practice to prevent vg call from being lost in giant snarls.
• --progress option added to vg call
• Paths with subranges on phase block 0 (or on haplotype paths without explicit phase blocks) can now be saved to GFA again.
• vg now supports reading from indexed GAF files of reads in many places where an indexed GAM file can be used

Updated Submodules

• htslib

vg 1.55.0 - Bernolda

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.55.0

Buildable Source Tarball: vg-v1.55.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• vg gamcompare now keeps track of whether reads could have possibly been gotten correct, based on there being a truth position for them in the truth set, and excludes reads without a truth from plots and accuracy statistics.
• vg giraffe no longer produces alignments with deletions on one end
• Multithreaded path cover / local haplotypes GBWT construction.
• CI should no longer fail due to truncated FTP download
• GBZ paper citation link now points to the right place
• vg surject will left align relative to the forward strand most of the time.

Updated Submodules

• gbwtgraph

vg 1.54.0 - Parafada

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.54.0

Buildable Source Tarball: vg-v1.54.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• Integrated haplotype sampling in vg giraffe now does diploid sampling.
• GBWTGraph algorithm for parsing GFA now handles P-line names of the form sample#contig correctly.

vg 1.53.0 - Valmontone

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.53.0

Buildable Source Tarball: vg-v1.53.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• vg sim no longer crashes on graphs with 1-node cycles
• vg autoindex can identify haplotypes specified as P-lines in a GFA
• Set reference samples in GBWT or GBZ with vg gbwt option --set-reference.
• vg rna no longer projects transcripts twice onto a reference given by RS tag in a GFA
• vg rna assigns unique names to twice-projected transcripts on cyclic haplotypes
• GBWT construction automatically increases buffer size if the paths are too long.
• In vg haplotypes, the default number of candidates for diploid sampling is now 32.
• vg giraffe now explains that --named-coordinates works for GAF output
• libvgio now uses quoted includes internally
• vg's README now prominently lists some recommended papers to cite when using parts of vg in your work
• Updated dozeu submodule should no longer crash vg giraffe and vg surject.

Updated Submodules

• dozeu
• gbwt
• libhandlegraph
• libvgio

vg 1.52.0 - Bozen

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.52.0

Buildable Source Tarball: vg-v1.52.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• vg construct now has a -A, --alt-paths-plain option for storing IDs from the VCF instead of hash-based IDs for alt allele paths.
• vg call patched so that certain problem cases no longer take forever.
• Mac CI now actually installs Node
• GBZ files can now hold reference paths like GRCh38#chr1, with no haplotype phase number.
• vg is now compatible with jq 1.7.
• Mac build should no longer fail with complaints about a missing atomic library.
• Tests should no longer fail due to odd alignments from diff.
• Better error messages from vg haplotypes.
• vg build process should now always use exactly one libhandlegraph
• add missing -O help for vg call
• vg Makefile now can take a CXX_STANDARD variable in. You should be able to e.g. make CXX_STANDARD=20 if you have a Protobuf/Abseil for C++20.
• GCSA2 construction in vg autoindex rewinds to pruning if memory is too high

Updated Submodules

• kff-cpp-api
• gcsa2
• libhandlegraph
• libvgio
• vcflib
• gbwtgraph

vg 1.51.0 - Quellenhof

Don't forget to mark the static binary executable:

chmod +x vg

Docker Image: quay.io/vgteam/vg:v1.51.0

Buildable Source Tarball: vg-v1.51.0.tar.gz

Includes source for vg and all submodules. Use this instead of Github's "Source Code" downloads; those will not build as they do not include code for bundled dependencies that the vg build process needs.

This release includes:

• Giraffe can do haplotype sampling automatically if sufficient inputs are provided.
• Simplified vg giraffe command line help; full list of options is still available with -h.
• Diploid mode for haplotype sampling: first select N haplotypes, then choose the best pair.
• Add ref-path stubbification option -S to vg clip
• vg validate now complains about duplicate path names
• vg CI expects only the allocated cores on the Gitlab runners
• vg CI Buildkit docker builds use the local Docker Hub mirror
• vg convert option --no-translation for converting GBWTGraph to GFA directly without using the node-to-segment translation.
• vg rna will not crash when adding transcripts with an intron of length 0
• vg paths now supports -H for selecting haplotype paths and -R for selecting reference paths

Updated Submodules

• backward-cpp
• gbwtgraph
• libbdsg