Changelog#

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased#

1.1.10 - 2026-01-16#

Fixed#

  • error with processing UniProt headers (#101)

  • error for no alignment result with MMseqs

  • critical error for selenoproteins is now an info (#102)

  • uppercase sequences before pyopal to avoid alphabet errors

1.1.9 - 2026-01-06#

Added#

  • --alignment-min-coverage parameter to filter low coverage alignments

  • generate-config CLI command - generate config file in case models are downloaded manually

  • calculate-maps CLI command - calculates all contact maps from PDB/MMCIF in directory (#100)

  • prediction matrices - predicted vectors allow to compare proteins with the help of distance analysis (#79)

  • custom scoring matrices - default VTML80

  • support for python < 3.13

Fixed#

  • Skipping extraction of broken structures from highquality_clust30, which caused segmentation fault (steineggerlab/foldcomp#56)

  • all output from the software log to stdout

  • np.genfromtext bug (#96)

  • contact map wrapping (#99)

  • query file losing sequences (#92)

  • MMseqs convertalis unlimited thread usage

1.1.8 - 2024-06-26#

Added#

  • fixed tests and added to CI/CD

  • added function to download MMseqs2 binaries for full CI/CD runs

  • removed Conda install - fully installable from PyPI

Fixed#

  • hetero atom bug introduced by Ruff (c044a158)

  • MacOS CI/CD bug connected to Foldcomp binary

1.1.7 - 2024-06-10#

Fixed#

  • resolved old naming issues (#84)

1.1.6 - 2024-06-04#

Added#

  • MMseqs2 API + tests

  • rudimentary documentation website

  • fixed bugs of v1.1.5

1.1.5 - 2024-04-01#

Added#

  • improved logging

  • support for ESM databases (#80)

  • corrected structures retrieval from PDB (#81, #83)

    • correct non-standard amino acids

    • remove base pairs

Changed#

  • retrieval of structures from PDB & FoldComp databases

  • optimized contact map alignment via Cython - improved search in large databases (ESM)

v1.1.4 - 2024-03-23#

Added#

  • --skip-pdb flag in case of unexpected PDB bugs

  • user parameter for allowed protein lengths --min-length and --max-length

Fixes#

  • installation of PDB database

  • alignment of faulty PDB structures

    • cases are reported via logging, require case-by-case investigation

    • faulty structures are then aligned to a predicted database, which are consistent

    • around 130/1300 alignments to PDB fail (10%)

v1.1.3 - 2024-03-13#

Fixes#

  • bug where MMseqs2 returned empty results search #73

v1.1.2 - 2024-02-22#

Added#

  • chunked loading of MMseqs2 preselected targets - improves runtime

Fixed#

  • C++17 lowered to C++14 for compatibility

  • pyopal API v0.5 integrated

  • removed scipy distance calculation with function

  • replaced biopython with biotite for structure manipulation better runtime and easier deploy

  • tests fixed & added to CI/CD

v1.1.1 - 2024-03-01#

Added#

  • Automatic setup of the PDB database during the first run

  • Automatic parsing of both mmCIF and PDB files

  • Removed torch dependency - significant improvement in startup times

  • added models of v1.1 - available for the download via get-models subcommand

  • CI/CD