overview of current rsnps functions

new-features.md

NCBI vs OPENSNP

• split ropensnp and ncbi functions? ¯_(ツ)_/¯
• adding a opensnp or dbsnp suffix to function name? see also https://ropensci.github.io/dev_guide/building.html#function-and-argument-naming

Function design

• what about searching for pos and chr?
• some of the functions only work for 1 SNP. extend for a vector of SNPs (e.g. genotypes())
• add progressbar

Add SNP info

• phenotypes
• eqtl
• https://www.ebi.ac.uk/gwas/

LD stuff

• http://grch37.rest.ensembl.org/
• https://github.com/statgen/LDServer
• previous ld_search function: https://github.com/ropensci/rsnps/commit/3109bd3238ee0ff136add1c22f02a1534cb2d718#diff-b686563661d51f11b33615acada264fb
• see also my blogpost.
• http://phase3browser.1000genomes.org/Homo_sapiens/Variation/HighLD?r=13:55312211-55313211;source=dbSNP;v=rs7981812;vdb=variation;vf=4837827
• https://ldlink.nci.nih.gov/?tab=ldmatrix

test-rsnps.R

## allgensnp -----------------------------

SNP1 <- "rs1528723" ## mine
SNP2 <- "rs7412" ## examples
SNP3 <- "rs143384"
dat1 <- allgensnp(snp = SNP1)
dat2 <- allgensnp(snp = SNP2)
dat3 <- allgensnp(snp = SNP3)

head(dat1)
head(dat2)
head(dat3)
# > table(dat2$local_genotype)
## CC   CT   TC   --   TT   00    T 
## 2807  356   66   43   19   16    1

## > head(dat3)
## name chromosome position                name   id genotype_id local_genotype
## 1 rs143384         20 35437976        R.M. Holston   22           8             AG
## 2 rs143384         20 35437976 Charles G. Sullivan 5326        3834             AA

## Improvements
## - better error output for allgensnp()
## - improve docu (what does 00 mean in the table above?)
## - columns names not described
## - what is "Curl options passed on to crul::HttpClient"
## - "value" description missing (see fetch_genotypes)
## - twice "name" as column 

## New features



## allphenotypes -----------------------------

dat <- allphenotypes(df = TRUE)

names(dat)
head(dat)
table(dat$characteristic)

## Improvements
## - not a dataframe, even with df = TRUE, but a list
## - columns names not described
## - number_of_users != length(unique(dat$id))
## - explain: phenotypes vs allphenotypes
## - "value" description missing (see fetch_genotypes)
## - dat %>% dplyr::filter(characteristic == "Height")

## New Feature

## annotations -------------------------------

annotations(snp = SNP3, output = "plos")
annotations(snp = SNP3, output = "mendeley")
annotations(snp = SNP3, output = "snpedia")
annotations(snp = SNP3, output = "metadata")

## Improvements
## - "value" description missing (see fetch_genotypes)

## New Feature
## - add feature eqtl
## - adding gwas catalog

## download_users ----------------------------

download_users(id = 33)
## wc -l ~/33.23andme.12.txt
download_users(id = "Sullivan") ## Sullivan is only a partial name

## Improvements
## - better error message for sullivan:  Error in strsplit(fileurl, "/")[[1]] : subscript out of bounds

## - value not very specific >> bc it downloads any file

## fetch_genotypes ----------------------------
data <- users(df = TRUE)
head( data[[1]] ) # users with links to genome data
mydata <- fetch_genotypes(url = data[[1]][1,"genotypes.download_url"],file="~/myfile.txt")

mydata
system("wc -l ~/myfile.txt")

## Improvements
## - difference between download_users and fetch_genotypes


## genotypes ---------------------------------

genotypes(SNP3, userid='1-20', df=TRUE)
genotypes(c(SNP3, SNP1), userid='1-20', df=TRUE)


## > head(dat3[,-4] %>% dplyr::arrange(id))
## name chromosome position id genotype_id local_genotype
## 1 rs143384         20 35437976  1           9             AG
## 2 rs143384         20 35437976  6           5             AG
## 3 rs143384         20 35437976  8           2             GG

## > genotypes(SNP3, userid='1-20', df=TRUE)
## snp_name snp_chromosome snp_position                 user_name user_id genotype_id genotype
## 1 rs143384             20     35437976 Bastian Greshake Tzovaras       1           9       AG
## 2 rs143384             20     35437976              Nash Parovoz       6           5       AG
## 3 rs143384             20     35437976         Samantha B. Clark       8           2       GG

## Improvements
## - difference between genotypes and allgensnp
## - better error message for: genotypes(c(SNP3, SNP1), userid='1-20', df=TRUE)
## - "value" description missing (see fetch_genotypes)


## ncbi_snp_query -----------------------------

## have another assembly than 38
ncbi_snp_query(c(SNP1, SNP2, SNP3))
ncbi_snp_query(c(SNP2))

# Query Chromosome    Marker Class  Gene Alleles Major Minor    MAF       BP AncestralAllele
# 1 rs1528723          8 rs1528723   snp UNC5D     A/T     A     T 0.0942 35269868     A,A,A,A,A,A
# 2    rs7412         19    rs7412   snp UNC5D     C/T     C     T 0.0751 35269868     C,C,C,C,C,C
# 3  rs143384         20  rs143384   snp UNC5D     C/T     C     T 0.4389 35269868     C,C,C,C,C,C

## -- wrong return of gene: 
#rs1528723 > UNC5D
#rs7412 > APOE
#rs143384 > GDF5

## THIS IS FIXED IN THE GH VERSION

## Improvements
## - difference between ncbi_snp_query2 and ncbi_snp_query
## - add pos.gr38
## - "value" description missing (see fetch_genotypes)



## ncbi_snp_query2 -----------------------------
ncbi_snp_query2(c(SNP1, SNP2, SNP3))
ncbi_snp_query2(c(SNP2))

## Improvements
## - no gene, but organism
## - "value" description missing (see fetch_genotypes)



## ncbi_snp_summary -----------------------------

ncbi_snp_summary(c(SNP1, SNP2, SNP3))

## Improvements

## - "value" description missing (see fetch_genotypes)
## - what is the gene2 column (gene:number)


## phenotypes -----------------------------------

phenotypes(userid='1-8', df=TRUE)

## Improvements

## - phenotypes vs phenotypes_byid vs allphenotypes
## - "value" description missing (see fetch_genotypes)


## phenotypes_byid --------------------------------

phenotypes_byid(phenotypeid = 57)