In khughitt/EuPathDB: Provides access to pathogen annotation resources available on EuPathDB databases

BiocStyle::markdown()

Authors: V. Keith Hughitt
Modified: r file.info("EuPathDB.Rmd")$mtime
Compiled: r date()

Overview

This tutorial describes how to query and make use of annotations from the EuPathDB : The Eukaryotic Pathogen Genomics Resource and the creation/access of local R packages derived from the EuPathDB.

Available EuPathDB Resources

The EuPathDB R package may also be used to generate local copies of some of the resources provided by eupathdb.org. These include:

1. BSgenome objects (raw genome sequences)
2. OrgDB objects (AnnotationDBI accessible tables of the genome/transcriptome annotations)
3. TxDB objects and the somewhat interchangeable GRanges objects. (transcript identifications)
4. OrganismDBI objects (AnnotationDBI tables with references to other databases like GO/Reactome/etc)

Because the EuPathDB is constantly updating their resources, it might be of use for one to generate these resources locally. In addition, the EuPathDB R package provides some shortcut functions for gathering data frames(tables) from these resources of annotation data.

Installing a new species in your R session

When AnnotationHub and the EupathDB web resources get out of sync, one might want to install and manipulate the newest version of the available Eupath data. The following blocks demonstrate how one might do those tasks.

Downloading and installing an arbitrary species

The following shows how one might gather information about Saccharomyces cerevisiae from fungidb.org.

An important caveat

If one is creating bsgenomes, then the number of open files may quickly pass the default limit of 1024 imposed by most linux systems. This will prove more than a little annoying, therefore you will very much want to do something like this:

```{bash ulimit, eval=FALSE} ulimit -HSn 4096 ## If you can do it, make this number higher.

On my host, I added the following to /etc/security/limits.conf

* soft nofile 81920

* hard nofile 409600

Which is admittedly overkill.

```r
library(EuPathDB)

## Ask for the version 42 data from fungidb for species with 'cerevisiae' in the name.
sc_entry <- get_eupath_entry(species="cerevisiae", webservice="fungidb", eu_version="v42")
sc_name <- sc_entry[["Species"]]
sc_entry

Creating the package

Now that we have the canonical name for yeast from fungidb, we can create a fresh orgdb package. I will not actually run this for the vignette because it takes a long time and prints a lot to screen.

orgdb_pkg <- make_eupath_orgdb(sc_entry)
txdb_pkg <- make_eupath_txdb(sc_entry)
bsgenome_pkg <- make_eupath_bsgenome(sc_entry)
organ_pkg <- make_eupath_organismdbi(sc_entry)

You will have to trust that the above worked in my R session and installed for me packages named:

1. OrgDB: org.Scerevisiae.S288c.v42.eg.db
2. OrganismDBI: fungidb.Saccharomyces.cerevisiae.S288c.v42
3. TxDB: TxDb.Saccharomyces.cerevisiae.S288c.FungiDB.v42
4. BSgenome: BSGenome.Saccharomyces.cerevisiae.S288c.v42
5. GRanges: An rda file named GRanges.Saccharomyces.cerevisiae.S288c.v42.rda

As the names suggest, these were derived from fungidb.org revision 43.

Extracting annotation data

The following demonstrates ways to extract data from the generated orgdb package.

Choosing and loading a generated package

Because the eupathdb is constantly evolving, the get_eupath_pkgnames() function will use the metadata generated in download_eupath_metadata() in order to provide a differently names package for each eupathdb revision.

orgdb_pkg <- get_eupath_pkgnames(sc_entry)
sc_orgdb <- orgdb_pkg$orgdb
## Here is the name of the current yeast package.
sc_orgdb
## Thus we see the v41 (as of late 2018), a number which presumably will continue increasing.
## We can set the version parameter to change this if we have a previous version installed.

## Now get the set of available columns from it:
library(sc_orgdb, character=TRUE)
pkg <- get0(sc_orgdb)
avail_columns <- AnnotationDbi::columns(pkg)
head(avail_columns)
## There are lots of columns!
length(avail_columns)

The EuPathDB provides quite an astonishing field of information. When creating the OrgDB packages, I prefixed column names of the various data types as follows:

1. ANNOT: The primary annotations. This includes a wide variety of information from descriptions to number of transmembrane domains.
2. GO: The Gene Ontology table.
3. ORTHOLOGS: COGs among EuPathDB species.
4. LINKOUT: Linkout entries provided by the EuPathDB.
5. INTERPRO: Interpro data from the eupathdb.
6. KEGGREST: Cross-references KEGG data using the R KEGGREST package.
7. PATHWAY: Pathway data from the eupathdb.
8. PUBMED: Pubmed entries related to each gene.

Extracting data from a package

Once we have a package loaded, everything else is an application of the AnnotationDbi interface. Here are a few examples.

The load_orgdb_annotation() and load_orgdb_go() are simply wrapper around AnnotationDbi in order to fill in the various arguments and more quickly return annotation of likely interest.

Full column list

1. gid: The Gene ID
2. annot_gene_name: Usually the genecard abbreviation for the gene.
3. annot_gene_type: Type of annotation, primarily 'protein coding'.
4. annot_cds: Raw coding sequence.
5. annot_cds_length: Length of the coding sequence. (oh, this is floating point now, that should be fixed)
6. annot_chromosome: Chromosome for this gene.
7. annot_ec_numbers: Enzyme Comission classification.
8. annot_ec_numbers_derived: EuPathDB defined EC.
9. annot_exon_count: Number of exons in this gene.
10. annot_five_prime_utr_length: Annotated length of the 5' UTR.
11. annot_gene_entrez_id: Cross reference ID to NCBI.
12. annot_gene_hts_nonsyn_syn_ratio: Ratio of known SNPs which result in synonymous mutations vs. non-synonymous mutations.
13. annot_gene_hts_nonsynonymous_snps: SNPs observed in this gene which make mutations that change the primary amino acid sequence.
14. annot_gene_hts_stop_codon_snps: SNPs observed in this gene which cause stop codons.
15. annot_gene_hts_synonymous_snps: SNPs observed in this gene which make mutations that do not change the primary amino acid sequence.
16. annot_gene_location_text: Textual encoding of the gene's location.
17. annot_gene_ortholog_number: Number of orthologs calculated by the eupathdb for this gene.
18. annot_gene_orthomcl_name: Name of the COG group calculated by the eupathdb.
19. annot_gene_paralog_number: Number of paralogs calculated by the eupathdb for this gene.
20. annot_gene_previous_ids: When gene IDs change in the eupathdb, the old IDs are recorded here.
21. annot_gene_product: Text description of the product of this gene.
22. annot_gene_source_id: ID of the source record for this gene.
23. annot_gene_total_hts_snps: Total number of SNPs observed for this gene.
24. annot_gene_transcript_count: How many transcripts are known for this gene?
25. annot_go_component: GO CC annotation.
26. annot_go_function: GO MF annotation.
27. annot_go_id_component: GO CC IDs, semicolon separated.
28. annot_go_id_function: GO MF IDs, semicolon separated.
29. annot_go_id_process: GO BP IDs, semicolon separated.
30. annot_go_process: GO BP annotation.
31. annot_has_missing_transcripts: Are there transcripts which are incomplete for this gene?
32. annot_interpro_description: InterPro description(s) for this gene.
33. annot_interpro_id: InterPro IDs or this gene.
34. annot_is_pseudo: Is this a pseudogene?
35. annot_isoelectric_point: Calculated isoelectric point of the product protein.
36. annot_location_text: Essentially a repeat of #16 above, odd.
37. annot_matched_result: I don't know what this is, I will filter it out soon.
38. annot_molecular_weight: Dalton weight of the protein product.
39. annot_organism: Redundant column telling what organism this is, it should be filtered.
40. annot_pfam_description: pFam annotations from the eupathdb.
41. annot_pfam_id: pFam IDs from eupathdb.
42. annot_pirsf_description: PIRSF annotations from the eupathdb.
43. annot_pirsf_id: PIRSF IDs from the eupathdb.
44. annot_predicted_go_component: CC annotation from Blast2GO provided by the eupathdb.
45. annot_predicted_go_function: MF annotation from Blast2GO provided by the eupathdb.
46. annot_predicted_go_id_component: CC IDs from Blast2GO from the eupathdb.
47. annot_predicted_go_id_function: MF IDs from Blast2GO.
48. annot_predicted_go_id_process: BP IDs from Blast2GO.
49. annot_predicted_go_process: BP annotations from Blast2GO.
50. annot_project_id: Name of the service, it should be filtered.
51. annot_prositeprofiles_description: Description from Prosite.
52. annot_prositeprofiles_id: ID from Prosite.
53. annot_protein_length: CDS length above / 3.
54. annot_protein_sequence: Primary amino acid sequence for this entry.
55. annot_sequence_id: I think this might be the chromosome ID?
56. annot_signalp_peptide: List of hits provided by signalP.
57. annot_signalp_scores: List of scores provided by signalP.
58. annot_smart_description: Descriptions provided by http://smart.embl-heidelberg.de/
59. annot_smart_id: Corresponding IDs from SMART.
60. annot_source_id: I think this one is also redundant.
61. annot_strand: forward/reverse?
62. annot_superfamily_description: Description provided by superfamily.
63. annot_superfamily_id: ID provided by superfamily.
64. annot_three_prime_utr_length: UTR following the CDS.
65. annot_tigrfam_description: Description provided by TIGRfam.
66. annot_tigrfam_id: IDs provided by TIGRfam.
67. annot_tm_count: Number of transmembrane domains annotated for this.
68. annot_trans_found_per_gene_internal: TM domains annotated inside the cell.
69. annot_transcript_index_per_gene: I need to look this up.
70. annot_transcript_length: Length from start to stop.
71. annot_transcript_link: I think this should be filtered.
72. annot_transcript_product: Text description for this transcript.
73. annot_transcript_sequence: Primary sequence from start to stop.
74. annot_transcripts_per_gene: How many annotated transcripts does this gene have?
75. annot_uniprot_id: UniProt database ID.
76. annot_wdk_weight: This should be filtered.
77. chr_id: Chromosome for each gene.
78. go_evidence_code: From the GO table, evidence codes, one entry/row.
79. go_id: GO ID, one entry/row.
80. go_is_not: Uncertain, likely should be filtered.
81. go_ontology: MF/BP/CC
82. go_reference: I don't recall.
83. go_sort_key: This should be filtered.
84. go_source: GO source code for each annotation.
85. go_support_for_evidence_code_assignment: Evidence code for each annotation.
86. go_term_name: Text name for each GO annotation.
87. go_transcript_id_s: This should be filtered I think.
88. interpro_description: From the INTERPRO table, descriptions, one entry/row.
89. interpro_e_value: E-value provided by interpro.
90. interpro_end_min: Score from interpro.
91. interpro_id: IDs from interpro, one entry/row.
92. interpro_name: Names for each family from interpro.
93. interpro_primary_id: Interpro primary IDs.
94. interpro_secondary_id: Interpro secondary IDs.
95. interpro_start_min: Unsure.
96. interpro_transcript_id_s: Transcripts associated with each interpro ID.
97. keggrest_kegg_geneid: KEGGREST table IDs from KEGG.
98. keggrest_ncbi_geneid: KEGGREST table IDs from NCBI.
99. keggrest_ncbi_proteinid: KEGGREST protein IDs.
100. keggrest_pathways: pathway IDs from KEGG.
101. keggrest_uniprotid: crossreferencing kegg to uniprot.
102. linkout_database: Linkout ID type, table of 1 entry/row.
103. linkout_ext_id: ID in the external linkout database.
104. linkout_source_id: Corresponding linkout ID.
105. orthologs_gid: gene ID in the orthologs table, one row/entry.
106. orthologs_organism: Species ID for this entry.
107. orthologs_product: Name of this ortholog group's product.
108. orthologs_syntenic: Are these genes syntenic in their locations?
109. pathway_ec_number_matched_in_pathway: The EC numbers in the pathway table, 1 record/row.
110. pathway_exact_ec_number_match: Do the EC numbers match?
111. pathway_expasy_url: Link to the pathway entry.
112. pathway_id: ID for this particular entry.
113. pathway_reactions_matching_ec_number: what it says on the tin.
114. pathway_source: Not sure.
115. pathway_source_id: Also not sure.
116. pubmed_authors: Authors of each paper/gene in pubmed, 1 paper/row.
117. pubmed_doi: DOI entries for the various papers.
118. pubmed_id: NCBI IDs for the papers.
119. pubmed_title: Title for each paper by gene.

## The columns which begin with strings like 'PATHWAY' or 'INTERPRO' are actually separate
## sql tables in the orgdb database, and as such will lead to a hugely redundant data table
## if we select them.
chosen_columns_idx <- grepl(x=avail_columns, pattern="^ANNOT")
chosen_columns <- avail_columns[chosen_columns_idx]

## Now we have a set of columns of interest, let us get a data table/data frame.
sc_annot <- load_orgdb_annotations(orgdb=sc_orgdb, keytype="gid", fields=chosen_columns)
## load_orgdb_annotations will fill out separate dataframes for each annotation type,
## genes, exons, transcripts, etc.  In this case, we only want the genes
## (The eupathdb does not provide much information for the others.)
sc_genes <- sc_annot[["genes"]]
dim(sc_genes)
head(sc_genes)
## Yay! We have data about S. cerevisiae!

chosen_columns_idx <- grepl(x=avail_columns, pattern="^GO")
chosen_columns <- avail_columns[chosen_columns_idx]
sc_go <- load_orgdb_go(sc_orgdb, columns=chosen_columns)
head(sc_go)
## Yay Gene ontology data for Crithidia!

chosen_columns_idx <- grepl(x=avail_columns, pattern="^INTERPRO")
chosen_columns <- avail_columns[chosen_columns_idx]
sc_interpro <- load_orgdb_go(sc_orgdb, columns=chosen_columns)
head(sc_interpro)
## Interpro data for Crithidia!

chosen_columns_idx <- grepl(x=avail_columns, pattern="^PATHWAY")
chosen_columns <- avail_columns[chosen_columns_idx]
sc_path <- load_orgdb_go(sc_orgdb, columns=chosen_columns)
head(sc_path)

Shortcut methods

Remembering the keytype and package names and everything can be annoying. The following attempts to make that easier. In this instance, the only thing we need to provide is a unique substring for the species of interest. If the substring is not unique, this should show the matches and choose the first.

## The function load_eupath_annotations() provides a shortcut to the above.
sc_annot <- load_eupath_annotations(species="S288c", eu_version="v42", webservice="fungidb")
dim(sc_annot)

Shortcut orthologs

A task I find myself needing to do fairly often is to get the set of genes orthologous to a given gene. There are lots of methods to handle this question; one of them includes the nice table provided by the eupathdb. Let us query that.

sc_ortho <- extract_eupath_orthologs(sc_orgdb)
dim(sc_ortho)
head(sc_ortho)
summary(sc_ortho)

Session Information

sessionInfo()

khughitt/EuPathDB documentation built on Nov. 4, 2023, 4:19 a.m.