In neurogenomics/orthogene: Interspecies gene mapping

orthogene: Interspecies gene mapping

orthogene is an R package for easy mapping of orthologous genes across hundreds of species.
It pulls up-to-date interspecies gene ortholog mappings across 700+ organisms.

It also provides various utility functions to map common objects (e.g. data.frames, gene expression matrices, lists) onto 1:1 gene orthologs from any other species.

In brief, orthogene lets you easily:

• convert_orthologs between any two species
• map_species names onto standard taxonomic ontologies
• report_orthologs between any two species
• map_genes onto standard ontologies
• aggregate_mapped_genes in a matrix.
• get all_genes from any species

Installation

if (!requireNamespace("BiocManager", quietly = TRUE))
     install.packages("BiocManager")
# orthogene is only available on Bioconductor>=3.14
if(BiocManager::version()<"3.14") BiocManager::install(version = "3.14")

BiocManager::install("orthogene")

library(orthogene)

data("exp_mouse")
# Setting to "homologene" for the purposes of quick demonstration.
# We generally recommend using method="gprofiler" (default).
method <- "homologene"  

Examples

Convert orthologs

convert_orthologs is very flexible with what users can supply as gene_df, and can take a data.frame/data.table/tibble, (sparse) matrix, or list/vector containing genes.

Genes, transcripts, proteins, SNPs, or genomic ranges will be recognised in most formats (HGNC, Ensembl, RefSeq, UniProt, etc.) and can even be a mixture of different formats.

All genes will be mapped to gene symbols, unless specified otherwise with the ... arguments (see ?orthogene::convert_orthologs or here for details).

Note on non-1:1 orthologs

A key feature of convert_orthologs is that it handles the issue of genes with many-to-many mappings across species. This can occur due to evolutionary divergence, and the function of these genes tends to be less conserved and less translatable. Users can address this using different strategies via non121_strategy=:

1. "drop_both_species" : Drop genes that have duplicate mappings in either the input_species or output_species, (DEFAULT).
2. "drop_input_species" : Only drop genes that have duplicate mappings in input_species.
3. "drop_output_species" : Only drop genes that have duplicate mappings in the output_species.
4. "keep_both_species" : Keep all genes regardless of whether they have duplicate mappings in either species.
5. "keep_popular" : Return only the most "popular" interspecies ortholog mappings. This procedure tends to yield a greater number of returned genes but at the cost of many of them not being true biological 1:1 orthologs.

When gene_df is a matrix. These strategies can be used together with agg_fun. This feature automatically performs both ortholog aggregation (many:1 mappings) and expansion (1:many mappings) of matrices, depending on the situation. This means that you have the option to keep non-1:1 ortholog genes, and still produce a matrix with only 1 gene per row. Options include: 1. "sum" 2. "mean" 3. "median" 4. "min" 5. "max"

For more information on how orthogene performs matrix aggregation/expansion, see the documentation for the underlying function: ?orthogene:::many2many_rows

gene_df <- orthogene::convert_orthologs(gene_df = exp_mouse,
                                        gene_input = "rownames", 
                                        gene_output = "rownames", 
                                        input_species = "mouse",
                                        output_species = "human",
                                        non121_strategy = "drop_both_species",
                                        method = method) 

knitr::kable(as.matrix(head(gene_df)))

Map species

map_species lets you standardise species names from a wide variety of identifiers (e.g. common name, taxonomy ID, Latin name, partial match).

All exposed orthogene functions (including convert_orthologs) use map_species under the hood, so you don't have to worry about getting species names exactly right.

You can check the full list of available species by simply running map_species() with no arguments, or checking here.

species <- orthogene::map_species(species = c("human",9544,"mus musculus",
                                              "fruit fly","Celegans"), 
                                  output_format = "scientific_name")
print(species)

Report orthologs

It may be helpful to know the maximum expected number of orthologous gene mappings from one species to another.

ortholog_report generates a report that tells you this information genome-wide.

orth_zeb <- orthogene::report_orthologs(target_species = "zebrafish",
                                        reference_species = "human",
                                        method_all_genes = method,
                                        method_convert_orthologs = method) 
knitr::kable(head(orth_zeb$map))
knitr::kable(orth_zeb$report)

Map genes

map_genes finds matching within-species synonyms across a wide variety of gene naming conventions (HGNC, Ensembl, RefSeq, UniProt, etc.) and returns a table with standardised gene symbols (or whatever output format you prefer).

genes <-  c("Klf4", "Sox2", "TSPAN12","NM_173007","Q8BKT6",9999,
             "ENSMUSG00000012396","ENSMUSG00000074637")
mapped_genes <- orthogene::map_genes(genes = genes,
                                     species = "mouse", 
                                     drop_na = FALSE)
knitr::kable(head(mapped_genes))

Aggregate mapped genes

aggregate_mapped_genes does the following:

1. Uses map_genes to identify within-species many-to-one gene mappings (e.g. Ensembl transcript IDs ==> gene symbols). Alternatively, can map across species if output from map_orthologs is supplied to gene_map argument (and gene_map_col="ortholog_gene").
2. Drops all non-mappable genes.
3. Aggregates the values of matrix gene_df using "sum","mean","median","min" or "max".

Note, this only works when the input data (gene_df) is a sparse or dense matrix, and the genes are row names.

data("exp_mouse_enst") 
knitr::kable(tail(as.matrix(exp_mouse_enst)))

exp_agg <- orthogene::aggregate_mapped_genes(gene_df=exp_mouse_enst,
                                             input_species="mouse", 
                                             agg_fun = "sum")
knitr::kable(tail(as.matrix(exp_agg)))

Get all genes

You can also quickly get all known genes from the genome of a given species with all_genes.

genome_mouse <- orthogene::all_genes(species = "mouse", 
                                     method = method)

knitr::kable(head(genome_mouse))

Session Info

utils::sessionInfo()

neurogenomics/orthogene documentation built on Jan. 30, 2024, 4:44 a.m.