Resources

Annotation

The Ensembl public MySQL Servers

The following script gives information on genes from ENSEMBL as well as attributes (columns) that contains gene.

library(biomaRt)
listMarts()
mart <- useMart("ENSEMBL_MART_FUNCGEN")
listDatasets(mart)
mart <- useMart("ensembl")
listDatasets(mart)
ensembl <- useMart("ensembl", dataset="hsapiens_gene_ensembl", host="grch37.ensembl.org", path="/biomart/martservice")
attr <- listAttributes(ensembl)
attr_select <- c('ensembl_gene_id', 'chromosome_name', 'start_position', 'end_position', 'description', 'hgnc_symbol', 'transcription_start_site')
gene <- getBM(attributes = attr_select, mart = ensembl)
filter <- listFilters(ensembl)
searchFilters(mart = ensembl, pattern = "gene")

See also https://sites.google.com/site/jpopgen/wgsa for precompiled annotation. Alternatively, 

# GENCODE v19
url <- "ftp://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_19/gencode.v19.chr_patch_hapl_scaff.annotation.gtf.gz"
gtf <- rtracklayer::import(url)
gencode <- as.data.frame(gtf)

Biobanks

Catalog

EFO

https://www.ebi.ac.uk/efo/

Example code,

library(ontologyIndex)

id <- function(ontology)
{
  inflammatory <- grep(ontology$name,pattern="inflammatory")
  immune <- grep(ontology$name,pattern="immune")
  inf <- union(inflammatory,immune)
  list(id=ontology$id[inf],name=ontology$name[inf])
}
# GO
data(go)
goidname <- id(go)
# EFO
file <- "efo.obo"
get_relation_names(file)
efo <- get_ontology(file, extract_tags="everything")
length(efo) # 89
length(efo$id) # 27962
efoidname <- id(efo)
diseases <- get_descendants(efo,"EFO:0000408")
efo_0000540 <- get_descendants(efo,"EFO:0000540")
efo_0000540name <- efo$name[efo_0000540]
isd <- data.frame(efo_0000540,efo_0000540name)
save(efo,diseases,isd,efoidname,goidname, file="work/efo.rda")
write.table(isd,file="efo_0000540.csv",col.names=FALSE,row.names=FALSE,sep=",")
pdf("efo_0000540.pdf",height=15,width=15)
library(ontologyPlot)
onto_plot(efo,efo_0000540)
dev.off()

eQTLGen

http://www.eqtlgen.org/

MetaMapLite

https://metamap.nlm.nih.gov/MetaMapLite.shtml

MR-Base/OpenGWAS

OpenTargets

These reflects v4 using GraphQL, https://platform-docs.opentargets.org/data-access/graphql-api.

Our first example is from the document (except ENSG00000164308) whose output is as ERAP2.json.

#!/usr/bin/bash

# https://platform.opentargets.org/api

module load ceuadmin/R
Rscript -e '
  ERAP2 <- subset(pQTLdata::caprion,Gene=="ERAP2")
  ERAP2$ensGenes
  data <- jsonlite::fromJSON("ERAP2.json")
  diseases_data <- data$data$target$associatedDiseases$rows
  diseases_data <- tidyr::unnest(diseases_data, cols = everything(), names_sep = "_")
  write.table(diseases_data, file = "ERAP2.tsv", sep = "\t", row.names = FALSE, quote = FALSE)
'

In fact it is effectively done as follows,

library(httr)
library(jsonlite)
gene_id <- "ENSG00000164308"
query_string <- "
  query target($ensemblId: String!){
    target(ensemblId: $ensemblId){
      id
      approvedSymbol
      associatedDiseases {
        count
        rows {
          disease {
            id
            name
          }
          datasourceScores {
            id
            score
          }
        }
      }
    }
  }
"
base_url <- "https://api.platform.opentargets.org/api/v4/graphql"
variables <- list("ensemblId" = gene_id)
post_body <- list(query = query_string, variables = variables)
r <- httr::POST(url = base_url, body = post_body, encode = 'json')
if (status_code(r) == 200) {
  data <- iconv(r, "latin1", "ASCII")
  content <- jsonlite::fromJSON(data)
} else {
  print(paste("Request failed with status code", status_code(r)))
}

A Bash implementation is copied here

curl 'https://api.platform.opentargets.org/api/v4/graphql' \
     -H 'Accept-Encoding: gzip, deflate, br' \
     -H 'Content-Type: application/json' \
     -H 'Accept: application/json' \
     -H 'Connection: keep-alive' \
     -H 'DNT: 1' \
     -H 'Origin: https://api.platform.opentargets.org' \
     --data-binary '{"query":"query targetInfo {\n  target(ensemblId: \"ENSG00000164308\") {\n    id\n    approvedSymbol\n    biotype\n    geneticConstraint {\n      constraintType\n      exp\n      obs\n      score\n      oe\n      oeLower\n      oeUpper\n    }\n    tractability {\n      label\n      modality\n      value\n    }\n  }\n}\n"}' \
     --compressed

The Python script can be used directly without change

import requests
import json

gene_id = "ENSG00000164308"
query_string = """
  query target($ensemblId: String!){
    target(ensemblId: $ensemblId){
      id
      approvedSymbol
      biotype
      geneticConstraint {
        constraintType
        exp
        obs
        score
        oe
        oeLower
        oeUpper
      }
      tractability {
        label
        modality
        value
      }
    }
  }
"""
variables = {"ensemblId": gene_id}
base_url = "https://api.platform.opentargets.org/api/v4/graphql"
r = requests.post(base_url, json={"query": query_string, "variables": variables})
print(r.status_code)
api_response = json.loads(r.text)
print(api_response)

Lastly we turn to R, which is necessary to get around httr::content(r) for iconvlist() with iconv().

library(httr)
library(jsonlite)

gene_id <- "ENSG00000164308"
query_string = "
  query target($ensemblId: String!){
    target(ensemblId: $ensemblId){
      id
      approvedSymbol
      biotype
      geneticConstraint {
        constraintType
        exp
        obs
        score
        oe
        oeLower
        oeUpper
      }
      tractability {
        label
        modality
        value
      }
    }
  }
"
base_url <- "https://api.platform.opentargets.org/api/v4/graphql"
variables <- list("ensemblId" = gene_id)
post_body <- list(query = query_string, variables = variables)
r <- httr::POST(url=base_url, body=post_body, encode='json')
data <- iconv(r, "latin1", "ASCII")
content <- jsonlite::fromJSON(data)

rentrez

The relevant URLs are as follows,

with example code,

library(rentrez)
entrez_dbs()
entrez_db_links("pubmed")
pubmed_fields <- entrez_db_searchable("pubmed")
# set_entrez_key("")
Sys.getenv("ENTREZ_KEY")
term <- "pQTLs OR (protein AND quantitative AND trait AND loci) AND human [MH] AND (plasma OR Serum)"
r <- entrez_search(db="pubmed",term=term,use_history=TRUE)
class(r)
names(r)
with(r,web_history)
unlink(paste("pubmed",c("fetch","summary"),sep="."))
fields <- c("uid", "pubdate", "sortfirstauthor", "title", "source", "volume", "pages")
for(i in seq(1,with(r,count),50))
{
  cat(i+49, "records downloaded\r")
  f <- entrez_fetch(db="pubmed", web_history=with(r,web_history), rettype="text", retmax=50, retstart=i)
  write.table(f, col.names=FALSE, row.names=FALSE, file="pubmed.fetch", append=TRUE)
  s <- entrez_summary(db="pubmed", web_history=with(r,web_history), rettype="text", retmax=50, retstart=i)
  e <- extract_from_esummary(s, fields)
  write.table(t(e), col.names=FALSE, row.names=FALSE, file="pubmed.summary", append=TRUE, sep="\t")
}
id <- 600807
upload <- entrez_post(db="omim", id=id)
asthma_variants <- entrez_link(dbfrom="omim", db="clinvar", cmd="neighbor_history", web_history=upload)
asthma_variants
snp_links <- entrez_link(dbfrom="clinvar", db="snp", web_history=asthma_variants$web_histories$omim_clinvar, cmd="neighbor_history")
all_links <- entrez_link(dbfrom='pubmed', id=id, db='all')

Roadmap

http://www.roadmapepigenomics.org/

snakemake workflow catalogue

https://snakemake.github.io/snakemake-workflow-catalog/

TWAS