Skip to content

Releases: ComparativeGenomicsToolkit/cactus

Cactus 2.7.1 2024-01-19

19 Jan 20:14
1771495
Compare
Choose a tag to compare

Cactus 2.7.1 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

This release adds some options to tune outgroup selection, as well as updates many included dependencies and tools

  • Add --chromInfo option to specify sex chromosomes of input genomes, in order to make sure outgroups are selected accordingly
  • Add --maxOutgroups option so that the number of outgroups can be toggled via the command line (previously required using a modified configuration file).
  • Update to Toil 6.0.0
  • Update to vg 1.54.0
  • Update to odgi 0.8.4
  • Update to latest taffy (fixing bug in paf export)
  • Update to abPOA 1.5.1
  • Fix Dockerfile so that Phast binaries are included
  • --indexMemory now acts as upper limit on chromosome-level jobs.

Cactus 2.7.0 2023-12-05

05 Dec 23:51
48410bd
Compare
Choose a tag to compare

Cactus 2.7.0 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

This release changes how outgroups are used during chaining during progressive alignment, and adds some pangenome options

  • Add --xg option for xg pangenome output.
  • Add (experimental) cactus-pagnenome --noSplit option in order to bypass reference chromosome splitting. This was previously only possible by running step-by-step and not using cactus-graphmap-join.
  • Add pangenome tutorial (developed for recent hackathon) to documentation.
  • Update to vg version 1.53.0.
  • Updated local alignment selection criteria. At each internal node of the guide tree Cactus picks a set of pairwise local alignments between the genomes being aligned to construct an initial sequence graph representing the whole genome alignment. This sequence graph is then refined and an ancestral sequence inferred to complete the alignment process for the internal node. The pairwise local alignments are generated with LASTZ (or SegAlign if using the GPU mode). To create a reliable subset of local alignments Cactus employs a chaining process that organizes the pairwise local alignments into pairwise chains of syntenic alignments, using a process akin to the chains and nets procedure used by the UCSC Browser. Previously, each genome being aligned, including both ingroup and outgroup genomes, was used to select a set of primary chains. That is, for each genome sequence non-overlapping chains of pairwise alignments were chosen, each of which could be to any of the other genomes in the set. Only these primary chains were then fed into the Cactus process to construct the sequence graph. This heuristic works reasonably well, in effect it allows each subsequence to choose a sequence in another genome with which it shares a most recent common ancestor. In the new, updated version we tweak this process slightly to avoid rare edge cases. Now each sequence in each ingroup genome picks primary chains only to other ingroup genomes. Thus the set of primary chains for ingroup genomes does not include any outgroup alignments. The outgroup genomes then get to pick primary chains to the ingroups, effectively voting on which parts of the ingroups they are syntenic too. The result of this change is that the outgroups are effectively only used to determine ancestral orderings and do not ever prevent the syntenic portions of two ingroups from aligning together.

Cactus 2.6.13 2023-11-15

15 Nov 15:12
d6fa7c4
Compare
Choose a tag to compare

Cactus 2.6.13 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

This release fixes an issue where Toil can ask for way too much memory for minigraph construction

  • Cut default minigraph construction memory estimate by half
  • Add --mgMemory option to override minigraph construction memory estimate no matter what
  • Exit with a clear error message (instead of more cryptic crash) when user tries to run container binaries in a container
  • Fix double Toil delete that seems to cause fatal error in some environments
  • Fix gfaffix regular expression bug that could cause paths other than the reference to be protoected from collapse.

Cactus 2.6.12 2023-11-07

07 Nov 18:23
368b8df
Compare
Choose a tag to compare

Cactus 2.6.12 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

The release fixes some recent regressions:

  • Include more portable (at least on Ubuntu) gfaffix binary.
  • Fix error where gpu support on singularity is completely broken.
  • Fix export_hal and export_vg job memory estimates when --consMemory not provided.

Cactus 2.6.11 2023-10-31

31 Oct 13:26
fb04f4f
Compare
Choose a tag to compare

Cactus 2.6.11 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

This release fixes a bug introduced in v2.6.10 that prevents diploid samples from working with cactus-pangenome

  • Remove stray assert False from diploid mash distance that was accidentally included in previous release

Cactus 2.6.10 2023-10-30

30 Oct 18:47
bd84de8
Compare
Choose a tag to compare

Cactus 2.6.10 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

This release contains bug fixes for MAF export and the pangenome pipeline

  • Patch taffy to fix bug where sometimes length fields in output MAF can be wrong when using cactus-hal2maf --filterGapCausingDupes
  • Fix regression in cactus-graphmap-split / cactus-pangenome so that small poorly aligned reference contigs (like some tiny unplaced GRCh38 contigs) do not get unintentionally filtered out. These contigs do not help the graph in any way, but the tool should do what it says and make a component for every single reference contig no matter what, which it is now fixed to do.
  • Cactus will now terminate with a clear error message if any other --batchSystem than single_machine is attempted from inside a docker container.
  • Mash distance order into minigraph construction fixed so that haplotypes from the same sample are always added contiguously in the presence of ties.
  • CI fixed to run all hal tests, and not just a subset.
  • pip install wheel added to installation instructions, as apparently that's needed to install Cactus with some (newer?) Pythons.

Cactus 2.6.9 2023-10-20

20 Oct 16:56
0c56e78
Compare
Choose a tag to compare

Cactus 2.6.9 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

This release contains some bug fixes and changes to docker image uploads

  • GFAffix updated to latest release
  • CI no longer pushes a docker image to quay.io for every single commit.
  • CPU docker release now made locally as done for GPU
  • --binariesMode docker will automatically point to release image (using GPU one as appropriate)
  • --consMemory overrides hal2vg memory as well
  • --defaulMemory defaults to 4Gi when using docker binaries
  • SegAlign job memory specification increased to something more realistic
  • --lastzMemory option added to override SegAlign memory -- highly recommended on SLURM
  • chromosome (.vg / .og) outputs from pangenome pipeline will have ref paths of form GRCh38#0#chr1 instead of GRCh38#chr1 to be more consistent with full-genome indexes (and PanSN in general)

Cactus 2.6.8 2023-09-28

28 Sep 16:29
ec14dd6
Compare
Choose a tag to compare

Cactus 2.6.8 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

This release includes several bug fixes for the pangenome pipeline

  • Fix bug where mash distances used to determine minigraph construction order could be wrong when input sample names differ by only their last character
  • Fix some job memory specifications to better support slurm environments
  • Guarantee that pangenome components have exactly two tips (one for each reference path endpoint). This is required for vg's now haplotype sampling logic.
  • Add warning message when genomes that are too diverse are input to cactus-pangenome
  • Update hal
  • --consMemory option now overrides memory for hal export, in addition to cactus_consolidated
  • Update vg to v1.51.0
  • Fix bug where samples passed in to --reference (except the first) could be dropped as reference in the final output if they are missing from the first chromosome
  • Port CI to OpenStack

Cactus 2.6.7 2023-08-16

17 Aug 00:03
b7c7b2e
Compare
Choose a tag to compare

Cactus 2.6.7 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

This release includes a patched vg and gfaffix

  • Update to vg version 1.50.1 which patches the path name incompatibility bug in 1.50.0
  • Revert minigraph-cactus Reference path name convention introduced in v2.6.6 (ie haplotypes can be left unspecified)
  • Upgrade to GFAffix 0.1.5 which fixes a crash among other things (thanks @danydoerr!)
  • Fix bug where large input contig sizes (>2Gb) would break the pangenome pipeline with some versions of awk.

Cactus 2.6.6 2023-08-03

03 Aug 13:35
d39b842
Compare
Choose a tag to compare

Cactus 2.6.6 is available in the following forms:

WARNING: do not use the github automatically generated source files (Source code (zip) or Source code (tar.gz)), these are not correct.

The Docker images and binaries linked above are built using AVX2 extensions, and require a CPU that supports them, except the "Pre-compiled Binaries For Older CPU Architectures" which should be compatible with any 64-bit architecture (and, since version 2.3.1, support Cactus's pangenome pipeline).

Please subscribe to the cactus-announce low-volume mailing list to receive notice of Cactus release.

Release Notes

This release fixes a compatibility problem between Cactus and the newly released vg version 1.50.0.

  • Update to vg v1.50.0
  • Patch hal2vg to never write reference path names of the form SAMPLE#CONTIG, as vg now fails with an exception when reading them with convert. Instead, SAMPLE#HAPLOTYPE#CONTIG is always used, even if there is no haplotype specified (in which case it's set to 0).
  • Add (prototype) --haplo option to build new subsampling-compatible giraffe indexes.