NEWS.md

In acidgenomics/AcidGenomes: Acid Genomics Genome Annotations

Release notes

AcidGenoems 0.7.2 (2023-12-06)

Minor changes:

• makeGRangesFromGff: Don't error on Seqinfo download failure from Ensembl via getChromInfoFromEnsembl. This is fragile for old GFF files currently, such as "Mus_musculus.GRCm38.90.gtf.gz". We have filed a bug report with Bioconductor for GenomeInfoDb, but we may need to write our own variant of this function that provides better support for old Ensembl releases. Consider referring to ensembldb code for an alternate approach.

AcidGenomes 0.7.1 (2023-12-05)

Minor changes:

• makeGRangesFromEnsembl: Fixed support for legacy GRCh38 release 87 dataset, which doesn't have gene versions defined, and would otherwise error if ignoreVersion = FALSE is set.

AcidGenomes 0.7.0 (2023-12-04)

Major changes:

• EnsemblToNcbi and NcbiToEnsembl: Added support for disabling strict mode, which is useful for mapping all genes in a reference genome. This is not a breaking change, as strict mode remains enabled by default. Also reworked the internal code to speed up 1:1 mapping return.

Minor changes:

• Hgnc: Renamed "symbol" column to "geneName" and "name" to "description", better matching naming conventions in other functions.

AcidGenomes 0.6.1 (2023-10-13)

Minor changes:

• Renamed HumanToMouse to JaxHumanMouse, which indicates the source (Jackson Laboratory) more clearly.
• Bug fixes for GENCODE and RefSeq genome downloads.

AcidGenomes 0.6.0 (2023-10-04)

New functions / classes:

• HumanToMouse: Downloads human-to-mouse gene mappings from the Jackson Laboratory MGI server. Returns with unique 1:1 mappings by default, which can be disabled with the unique argument.

Major changes:

• Renamed some identifier classes to use "To" instead of the numeric "2" in the class name. Sorry Prince, but this just looks weird for some functions. Applies to Ensembl2Ncbi, Gene2Symbol, Ncbi2Ensembl, Protein2Gene, and Tx2Gene.
• Hardened class checks via validObject for our custom classes, ensuring that metadata is consistently slotted, including date and package version.
• Switched to HTTPS downloads internally for some functions, as the Ensembl FTP server has been prone to timeouts lately. This does not affect our primary genome downloading functions, which still use FTP for speed.

Minor changes:

• Migrated internal data from inst/extdata to save internally inside the package via sysdata.rda file. Applies to detectOrganism and mapNcbiTaxId currently.

AcidGenomes 0.5.1 (2023-08-01)

Major changes:

• downloadGencodeGenome: This genome download function now sanitizes the transcriptome FASTA to only include transcript identifiers in the header without additional information separated by the "|" (pipe) delimeter. This approach is not commonly used in FASTA files, and results in unwanted downstream behavior when quantifying at transcript level using kallisto and minimap2. Note that salmon can currently handle this edge case when setting the --gencode flag during the genome index step. This action is non-destructive and returns a "transcriptome_fixed" FASTA file. We are now symlinking this fixed file by default, but the unmodified original is retained in the transcriptome download folder.

Minor changes:

• currentEnsemblGenomeBuild: Fixed internal REST API query to Ensembl server. This now requires "content-type=application/json" to be defined in the URL, otherwise the Ensembl server returns text instead of JSON.
• currentEnsemblVersion: Now parses current_README file on FTP server instead of the top level README. We changed this because the README symlink can break during Ensembl release updates (e.g. 109 to 110, in progress).

AcidGenomes 0.5.0 (2023-04-27)

Starting a new release series to denote potential breaking changes with legacy objects saved with entrezId instead of ncbiGeneId.

New functions:

• Added utility functions to easily map gene symbols to stable gene identifiers at Ensembl, HGNC, and NCBI (Entrez): mapGeneNamesToEnsembl, mapGeneNamesToHGNC, mapGeneNamesToNCBI. These are covered against Homo sapiens and Mus musculus.

Major changes:

• Classes that extend GRanges (e.g. EnsemblGenes, EnsemblTranscripts, GencodeGenes, GencodeTranscripts) now intentionally fail class checks if entrezId is defined instead of ncbiGeneId in mcols metadata. This makes downstream handoff to GSEA functions in AcidGSEA easier to manage. For legacy objects, use updateObject to resolve this check.
• makeGRangesFromEnsembl and makeGRangesFromGFF now attempt to fetch additional useful gene metadata, including gene synonyms from the Ensembl FTP server when applicable. This currently applies to gene annotation files from Ensembl and GENCODE. Note that extra metadata is not supported for legacy Homo sapiens GRCh37 genome build.

Minor changes:

• Renamed EntrezGeneInfo function to NcbiGeneInfo.
• Removed geneSynonyms from NAMESPACE. Consider using NcbiGeneInfo or HGNC instead for synonym information.
• Removed HGNC2Ensembl and MGI2Ensembl. Just use HGNC and MGI function return instead.
• downloadEnsemblGenome now downloads additional useful metadata files.
• Added updateObject support to update legacy objects that may fail new entrezId class checks.

AcidGenomes 0.4.8 (2023-02-16)

Minor changes:

• makeTx2GeneFromGFF, Tx2Gene: Ensure that rownames are defined for RefSeq genome annotations, which are constructed from GenomicRangesList method.

AcidGenomes 0.4.7 (2023-02-10)

New functions:

• gencodeReleaseHistory: This function scrapes the GENCODE website to return the full release history for either human or mouse genomes.

Minor changes:

• currentEnsemblVersion: Fix for breaking change on Ensembl FTP server. The file this function parses has been renamed from current_README to simply README.
• mapGencodeToEnsembl: Now using gencodeReleaseHistory internally to dynamically fetch metadata directly from the GENCODE website, instead of relying on an internal CSV mapping file. This helps avoid having to update the package every time a new Ensembl/GENCODE release comes out.

AcidGenomes 0.4.6 (2023-02-09)

Minor changes:

• Migrated requireNamespaces import from AcidBase to goalie.
• Updated dependency versions.
• Now allowing Ensembl Seqinfo fetch step to fail, to avoid unit test issues with Mus musulus Ensembl 90 GTF file (for bcbioRNASeq).

AcidGenomes 0.4.5 (2023-01-31)

Minor changes:

• downloadUCSCGenome: Added a manual override for defaulting to "hg38" for Homo sapiens, which has switched over to experimental "hs1" T2T genome build.
• mapGencodeToEnsembl: Added support for mapping Mus musculus releases.
• Internal Seqinfo generators now default to pulling annotations from Ensembl for GENCODE reference, rather than UCSC. Note that UCSC Seqinfo function is currently broken for hg38, but will be fixed in pending GenomeInfoDb v1.34.8.
• Updated unit tests to check against latest genome build releases.

AcidGenomes 0.4.4 (2022-10-25)

Minor changes:

• Compatibility update to provide support for breaking changes introduced with pipette 0.10.0 release.
• EntrezGeneInfo: Added column name checks. Renamed xTaxId to taxonomyId.
• Migrated all importFrom calls to imports.R file.
• Removed BiocParallel bplapply usage.
• export: Updated methods to match new conventions defined in pipette.
• Improved support for quick detection of NCBI taxonomic group for commonly used model organisms. Previously human and mouse were supported. Added support for zebrafish, rat, worm, fruitfly, and yeast.

AcidGenomes 0.4.3 (2022-06-09)

Minor chagnes:

• makeGRangesFromEnsembl: Hardened internal code to suppress spurious warnings from rtracklayer due to masking of download.file function. See issue #71 for details.
• Improved temporary directory creation and deletion using tempdir2 and unlink2 internally, which improves support for continuous integration (CI) checks on Windows.

AcidGenomes 0.4.2 (2022-05-27)

Minor changes:

• Ensembl2Entrez: Simplified format "1:1" handling, based on new code approach used in AcidGSEA package. This now keeps track of original rownames, which is necessary for bcbioRNASeq clusterProfiler R Markdown template.

AcidGenomes 0.4.1 (2022-05-24)

Minor changes:

• Updated lintr checks and testthat unit tests.

AcidGenomes 0.4.0 (2022-05-04)

Major changes:

• S4 classes containing identifier mappins now are set to contain DFrame instead of DataFrame, as the previous definition approach no longer works with Bioconductor 3.15. This applies to Ensembl2Entrez, Entrez2Ensembl, EntrezGeneInfo, Gene2Symbol, HGNC, HGNC2Ensembl, Protein2Gene, and Tx2Gene classes.
• downloadGencodeGenome: Added support for Entrez and RefSeq identifiers, which are now defined in mcols of GRanges objects.
• Reworked internal code of rtracklayer runner functions. In particular, improved identifier handling for RefSeq GFF and GTF runners, adding support for extraction of Entrez gene identifiers.

Minor changes:

• Reformatted code using styler conventions.
• Removed internal dependencies on stringr, in favor of base R when possible and stringi package otherwise.
• Hardened import of external files using readr package, where applicable. Applies primarily to Entrez, HGNC, and MGI file import.
• makeTx2GeneFromFASTA: Added ignoreVersions argument, which is now enabled by default, to match other Tx2Gene functions.
• Updated unit tests to cover most recent versions of genome assembly files.

AcidGenomes 0.3.0 (2022-03-11)

New functions:

• mapGencodeToEnsembl: Convenience function for mapping human GENCODE release (e.g. 39) to corresponding Ensembl release (e.g. 105).

Minor changes:

• Reworked NAMESPACE to inherit from Bioconductor packages, migrating from previous approach that used generics defined only in AcidGenerics.
• Relaxed GenomicRanges mcols column name validity checks to no longer require strict camelCase formatting. This check can run into issues when slotting into a DESeqDataSet object, which appends non-camelCase-formatted columns into the mcols (corresponding to the rowRanges of the object).
• Migrated some generics to AcidGenerics, and reexporting here (non-breaking change): Ensembl2Entrez, Entrez2Ensembl, Gene2Symbol, Tx2Gene.
• Updated references from GRanges to GenomicRanges (virtual class) where applicable in the package.
• Split out documentation on Ensembl2Entrez and Entrez2Ensembl methods into separate files.
• Increased the verbosity of setMethod calls, particular in the signature argument, where applicable.
• Calls to AnnotationHub internally now should never prompt the user about whether to create cache directory. This is achieved by setting ask = FALSE internally.
• Updated internal class checks inside of assert calls to check for GenomicRanges virtual class rather than GRanges.
• Improved alphabetical sorting of exported S4 methods.
• Reexporting some additional functions from GenomeInfoDb: Seqinfo, genome, seqinfo, and seqlevels. Note that corresponding assignment methods (if applicable) are intentionally not reexported here.
• Genome download functions no longer cache using BiocFileCache by default.

AcidGenomes 0.2.20 (2022-01-08)

Minor changes:

• Tx2Gene class check: disabling check that looks for identical transcript and gene identifiers. This check is not compatible with the Saccharomyces cerevisiae (sacCer3) reference genome. Thanks for pointing this out @amizeranschi.

AcidGenomes 0.2.19 (2021-09-13)

Minor changes:

• getEnsDb / makeGRangesFromEnsembl: Quieted down package loading from Bioconductor when obtaining annotations for GRCh37 (EnsDb.Hsapiens.v75 release package).
• Improved CLI messages to use S4 class when applicable.

AcidGenomes 0.2.18 (2021-09-03)

Minor changes:

• Improved CLI message formatting.
• downloadRefSeqGenome, downloadUCSCGenome: Improved the genomeBuild documentation, with more specific examples.

AcidGenomes 0.2.17 (2021-08-11)

Major changes:

• Gene2Symbol: Hardened internal identifier mapping code in switch call to support format argument. Improved unit testing for expected behavior of format argument. Fixed "1:1" mapping to split based on geneName column rather than geneId column.
• Tx2Gene: Improved code coverge and cleaned up internal complete.cases handling.
• All identifier mapping classes (e.g. Ensembl2Entrez, Gene2Symbol, Tx2Gene) now check for complete.cases in S4 validity methods.

Minor changes:

• currentEnsemblVersion and mapHumanOrthologs working examples are now re-enabled, wrapped in a try call.

AcidGenomes 0.2.16 (2021-08-09)

Major changes:

• Removed mapEnsemblBuildToUCSC and mapUCSCBuildToEnsembl functions. Also removed mapping support for UCSC genome build names (e.g. "hg38") inside of makeGRangesFromEnsembl calls, since this is not technically the correct genome build name.
• downloadEnsemblGenome, downloadGencodeGenome, etc. now support file caching by default with cache = TRUE argument.

Minor changes:

• Improved genome download code coverage.
• Split out stripGeneVersions and stripTranscriptVersions documentation into separate files.
• Reorganized S4 method documentation to be alphabetical consistently.
• export: Hardened Tx2Gene method to ensure that rownames are consistently removed prior to export. Noticed that this was an issue with UCSC genome build download.

AcidGenomes 0.2.15 (2021-07-27)

Minor changes:

• Added support for new Mus musculus GRCm39 genome build. Updated internal code for mapUCSCBuildToEnsembl and downloadEnsemblGenome, in particular. Note that *_chr_patch_hapl_scaff GFF and GTF files are no longer available on the Egnyte FTP server for GRCm39 (only GRCm38 and GRCh38).

AcidGenomes 0.2.14 (2021-06-10)

Minor changes:

• Gene2Symbol: Improve handling when gene identifiers are integer, such as is the case with NCBI Entrez gene identifiers.

AcidGenomes 0.2.13 (2021-05-18)

• Internal fixes to provide compatibility for R 4.1 release.
• Updated dependencies to support Bioconductor 3.13.

AcidGenomes 0.2.12 (2021-04-27)

Major changes:

• Removed some Bioconductor packages from imports: AnnotationDbi, AnnotationHub, GenomeInfoDb, and ensembldb. This helps reduce package loading time and avoid unwanted BiocManager messages from appearing at startup (due to loading of AnnotationHub). These are included as suggested packages, which should not be problematic, since they are frequently used.

AcidGenomes 0.2.11 (2021-03-19)

Minor changes:

• HGNC now returns columns with split values as CharacterList, instead of as character strings containing "|".

AcidGenomes 0.2.10 (2021-03-15)

Minor changes:

• mapHumanOrthologs: Hardened mouse-to-human matching.
• makeGRangesFromEnsembl: No longer hard-coding minimum release version check at 87, in case older releases are ported to AnnotationHub in a future release.
• Revert back to ignoreVersion = TRUE by default for genome annotation importers, as this is typically what users expect by default.

AcidGenomes 0.2.9 (2021-03-03)

Minor changes:

• Gene2Symbol functions now preserve metadata, as expected. This was causing pointillism package to error, due to unwanted breaking change.
• Tx2Gene: Improved consistency of metadata return, ensuring call and synonyms are not defined.
• Renamed internal "acidGenomes" metadata key to "packageVersion", for consistency with conventions used in other Acid Genomics packages.

AcidGenomes 0.2.8 (2021-03-02)

Minor changes:

• Relaxed validity checks for EnsemblGenes and EnsemblTranscripts.

AcidGenomes 0.2.7 (2021-02-26)

Minor changes:

• makeGRangesFromGFF: Improved support and code coverage for handling of bcbio-nextgen ref-transcripts.gtf genome file.

AcidGenomes 0.2.6 (2021-02-25)

Minor changes:

• Updated basejump dependency versions.
• EntrezGeneInfo: Improved column formatting.

AcidGenomes 0.2.5 (2021-02-17)

Minor changes:

• Genome downloader functions (e.g. downloadEnsemblGenome) now return relative symlinks instead of absolute paths.
• Bug fix for mapHumanOrthologs internal join step. Now returns humanGeneId and humanGeneName columns instead of hgncId and hgncName columns, which technically were incorrect, since these map to Ensembl.

AcidGenomes 0.2.4 (2021-02-13)

New functions:

• EntrezGeneInfo: New utility for obtaining gene annotations from NCBI.

Major changes:

• geneSynonyms: Reworked internal code, extending EntrezGeneInfo.

AcidGenomes 0.2.3 (2021-02-10)

Minor changes:

• Reverted back to using vroom as importer for HGNC and MGI2Ensembl.

AcidGenomes 0.2.2 (2021-02-09)

Minor changes:

• Reduced the number of reexported functions.

AcidGenomes 0.2.1 (2021-02-08)

Minor changes:

• Now including some reexports from GenomicRanges and IRanges.

AcidGenomes 0.2.0 (2021-02-02)

New functions:

• Added new genome download functions, migrated from previous Python approach defined in py-koopa package.
• Switched from cli to AcidCLI for interactive messages.

Major changes:

• In metadata columns, renamed "ID" to "Id", for stricter lower camel case formatting.
• Now ensuring GenomicRanges gets attached as a dependency.
• Now using BiocFileCache (via pipette::cacheURL internally) to automatically cache GFF/GTF files when used in makeGRangesFromGFF.
• Renamed ignoreTxVersion to simply ignoreVersion, where applicable. We want this setting to also apply at gene level.
• GFF, TxDb, and ensembldb parser functions now default to ignoreVersion = FALSE. Previous releases of AcidGenomes and basejump had this set to ignoreVersion = TRUE by default. Note that both modes are now non-destructive.
• GRanges mcols now return with tx prefix instead of transcript.
• GRanges mcols now use strict camel case formatting (e.g. geneId instead of geneID).

AcidGenomes 0.1.1 (2020-10-12)

Minor changes:

• Relaxed stringency of internal organism name checks. Applies to makeGene2SymbolFromEnsembl for example, which is causing running examples to fail in pointillism without a fix.
• Updated minimum dependency versions.

AcidGenomes 0.1.0 (2020-10-07)

Initial release, consisting of functions migrated from basejump.

acidgenomics/AcidGenomes documentation built on Dec. 10, 2023, 10:35 p.m.