NEWS.md

In patzaw/BED: Biological Entity Dictionary (BED)

Version 1.5.1

Changes

• Correction of a bug in dumping Ensembl data
• Generalization of the function to download uniprot data

Version 1.5.0

Changes

• Support version 5 of neo4j graph DB.

Version 1.4.13

Changes

• Need to explicitly call connectToBed() after loading the package (this function is not automatically called on package load anymore to avoid warning message during CRAN checks)

Version 1.4.12

Bug fix

• compareBedInstances(): platform comparison

Version 1.4.10/11

Improvements

• Improved display of beidsModule and of findBeids() interfaces
• Default values for focusOnScope() parameters
• Minor adjustments in the vignette

Version 1.4.9

Bug fix

• Make sure that is.character(metadata(BEIDList)$.lname)

Rebuild BED

• Chose ftp site for dumpUniprotDb()

Version 1.4.8

Neo4j changes

• connectToBed() supports the .opts parameter of RCurl::curlPerform().

Version 1.4.7

Neo4j changes

• "CREATE UNIQUE" replaced by "MERGE" in neo4j queries

Version 1.4.6

Bug fix

• Suppress wrong warning in listBeIdSources()

Version 1.4.5

Improvements

• Use https protocol for getting NCBI data
• dumpNcbiTax: using the zip archive instead of the tar.gz because of internal issue with firewall.
• Correct canonical products from Ensembl

Version 1.4.4

Improvements

• convBeIds() and getBeIdConvTable() have an additional parameter "canonical" to only keep canonical symbols of BE when converting from or to symbols.

Rebuild BED

• dumpUniprotDb() uses less memory

Version 1.4.3

Complying with CRAN policies

• Software names in single quotes ('Neo4j' and 'Docker') in title and description fields of the DESCRIPTION file.
• Reference added in the description field of the DESCRIPTION file.
• Link to the 'Docker' image of the 'Neo4j' database added in the description field of the DESCRIPTION file.
• Author and contribution declared in the Authors@R field of the DESCRIPTION file.
• Functions forgetBedConnection(), dumpEnsCore(), dumpNcbiDb(), dumpNcbiTax(), dumpUniprotDb(), getEnsemblGeneIds(), getEnsemblPeptideIds(), getEnsemblTranscriptIds(), getNcbiGeneTransPep(), getUniprot(), loadNCBIEntrezGOFunctions(), and loadNcbiTax() functions do not write anymore by default in the user's home file space.
• As explained in the documentation and in code comments, query results are automatically written on user's home file space if and only if the cache parameter has been set to TRUE when calling the connectToBed() function (by default this parameter is set to FALSE).

Version 1.4.2

Bug fixes

• Correction of a bug when no data is returned by listBeIdSources
• Type correction before inner_join when getting data

Version 1.4.0

New features

• Shiny module: beidsServer() and beidsUI() functions
• New Shiny gadget (RStudio addin): findBeids()
• guessIdOrigin() (still available) has been renamed guessIdScope(). It takes into account user input to guess the identifiers scope.
• The scope of the identifiers to convert or to explore is automatically guessed when not provided by the user. It is still recommended to provide them when they are known but this feature could help for exploratory session.
• Find all identifiers in all scopes corresponding to BE:
  • beIDsToAllScopes() (more specific than geneIDsToAllScopes())

Implementation changes

• By default the system does not remember connections for policy reasons.
• By default the system does not use cache for policy reasons. However, it is recommended to set it to TRUE when connecting to improve the speed of recurent queries.
• merge() calls have been replaced by dplyr::inner_join().
• Import statements have been replaced as much as possible by explicit :: calls.

Version 1.3.0

New features

• Use of lucene indexes for fast and flexible searches of BE:
  • searchBeid()
• Find all identifiers in all scopes corresponding to genes:
  • geneIDsToAllScopes()

Implementation changes

• Simplification of queries for getting gene descriptions

Rebuild-BED.Rmd

• Not in the final package anymore
• Ensembl 100
• Based on neo4j-community-3.5.17
• GO molecular functions are not included anymore (more relevant tools should be used for that purpose)

Version 1.2.4

Implementation changes

• Correct a bug when chosing a be in searchId
• Avoid the application to crash when there is no running neo4j instance (This bug appeared with version 2 of neo2R)

Version 1.2.3

Rebuild-BED.Rmd

• Ensembl 99
• Based on neo4j-community-3.5.14
• NCBI does not support mapping to unigene and vega anymore. Unigene and Vega identifiers are not available anymore in BED as well.

Implementation changes

• Prepopulate nodes with a subset and use 'CALL db.resampleOutdatedIndexes();' to speed up the feeding process.

Version 1.2.2

Rebuild-BED.Rmd

• Ensembl 98
• Based on neo4j-community-3.5.11
• Because of a few (a very small number but quite annoying) dubious mapping with EntrezGene identifier, only unambigous mapping from Ensembl have been conserved. Relevant ambigous mapping be rescued by mapping from the NCBI and to other reference database such as HGNC, MGI or RGD. The cleanDubiousXRef() has been updated to this end Examples:
  • https://www.ensembl.org/Homo_sapiens/Gene/Matches?g=ENSG00000120088

Version 1.2.1

Implementation changes

• Improving mapping function for supporting difficult history of Zebrafish Ensembl identifiers

Version 1.2.0

Implementation changes

• cleanDubiousXRef() function applied to NCBI gene cross-references from Ensembl

Rebuild-BED.Rmd

• Based on neo4j-community-3.5.9
• Add Danio rerio (zebrafish) support

Version 1.1.9

Implementation changes

• Correction of findBe() result display
• Correction of bug in getBeIds() and getRelevantIds()
• Implement BEIDList class

Version 1.1.8 - Release notes - 2019-07-29

Implementation changes

• Correction of how Ensembl data files are pre-processed because of the apparition of "carriage return" (\r\n) characters.
• Give write rights on the neo4j database before copying it in the docker image.

Rebuild-BED.Rmd

• Based on neo4j-community-3.5.7
• Ensembl 97

Version 1.1.7 - Release notes - 2019-05-08

Implementation changes

• When associating one id to another, the BE identified by the first id is deleted after that its production edges have been transferred. After this operation all id "corresponding_to" the first id do not directly identify any BE as they should be supposed to do. Thus, it's not recommended to associate such ID to others.

Version 1.1.6 - Release notes - 2019-04-23

Implementation changes

• Split the toImport file in the bedImport function before loading csv in neo4j ==> improve robustness and efficiency of the load
• Separate import of nodes and import of edges to avoid neo4j eager issues
• Add Sus scrofa support in the dumpUniprotDb function
• Ensembl history has been adapted to avoid loop (found for some pig transcripts)

Rebuild-BED.Rmd

• Based on neo4j-community-3.5.4
• Ensembl 96
• Add GPL6887 platform
• Add Sus scrofa (pig) support
• Set the following cypher parameters during the building to avoid query replanning (https://neo4j.com/blog/cypher-write-fast-furious/):
  • --env=NEO4J_cypher_minreplaninterval=100000000ms \
  • --env=NEO4J_cypher_statisticsdivergencethreshold=1 \

Version 1.1.5 - Release notes - 2019-01-30

Methodological changes

• Improving case insentive conversion from and to symbols

Rebuild-BED.Rmd

• Based on neo4j-community-3.4.12
• Ensembl 95

Version 1.1.4 - notes

• Notes about converting from and to gene symbols in the BED vignette.
• Simplification of the feeding procedure based on docker (see README.md)

Version 1.1.3 - notes

Bug fixes

• Avoid duplicated lines returned by convBeIds when converting from Symbols.

Version 1.1.2 - Release notes - 2018-07-20

Methodological changes

• More accurate ortholog mapping by avoiding the use of deprecated gene identifiers in the convBeIds function.

Vignettes

• BED: based on bed-ucb-human:2018.07.20 (https://hub.docker.com/r/patzaw/bed-ucb-human/)

Rebuild-BED.Rmd

• Based on neo4j-community-3.4.4
• Ensembl 93

Bug fixes

• searchId: focusing on one or several BE

patzaw/BED documentation built on April 30, 2024, 5:31 a.m.