Takes variants as input and annotates them using borads oncotator api (http://www.broadinstitute.org/oncotator/). Output is a dataframe of annotated variants in maf format.

oncotate.r

# Usage
# 
# oncotate(maflite, header = FALSE, basename = NULL)
# Arguments
# 
# maflite	
# input tsv file with chr, start, end, ref_allele, alt_allele columns. (rest of the columns, if present will be attached to the output maf)
#
# header	
# logical. Whether input has a header line. Default is FALSE.
#
# basename	
# NULL. if basename is given, annotations will be written to <basename>.maf file.
#
# Details
# 
# Input should be a five column file with chr, start, end, ref_allele, alt_allele (and so on, but only first five will used, rest will be attached to resulting maf file). Note: Time consuming if input is huge.

oncotate = function (maflite, header = FALSE, basename = NULL) {
    require(package = "rjson")
    anno.df = c()
    m = read.delim(maflite, stringsAsFactors = F, header = header)
    anno = paste(m[, 1], m[, 2], m[, 3], m[, 4], m[, 5], sep = "_")
    for (i in 1:length(anno)) {
        rec = anno[i]
        rec.url = paste("http://www.broadinstitute.org/oncotator/mutation", 
            rec, sep = "/")
        annot = fromJSON(file = rec.url)
        anno.df = rbind(anno.df, as.data.frame(annot))
    }
    colnames(anno.df) = gsub(pattern = "^X", replacement = "", 
        x = colnames(anno.df))
    colnames(m)[1:5] = c("Chromosome", "Start_Position", "End_Position", 
        "Reference_Allele", "Tumor_Seq_Allele2")
    anno.df = cbind(m, anno.df)
    anno.df$Center = NA
    anno.df$Tumor_Seq_Allele1 = anno.df$Reference_Allele
    colnames(anno.df)[which(colnames(anno.df) == "gene")] = "Hugo_Symbol"
    colnames(anno.df)[which(colnames(anno.df) == "variant_classification")] = "Variant_Classification"
    colnames(anno.df)[which(colnames(anno.df) == "variant_type")] = "Variant_Type"
    colnames(anno.df)[which(colnames(anno.df) == "HGNC_Entrez.Gene.ID.supplied.by.NCBI.")] = "Entrez_Gene_Id"
    colnames(anno.df)[which(colnames(anno.df) == "strand")] = "Strand"
    colnames(anno.df)[which(colnames(anno.df) == "build")] = "NCBI_Build"
    colnames(anno.df)[which(colnames(anno.df) == "strand")] = "Strand"
    anno.df1 = anno.df[, c("Hugo_Symbol", "Entrez_Gene_Id", "Center", 
        "NCBI_Build", "Chromosome", "Start_Position", "End_Position", 
        "Strand", "Variant_Classification", "Variant_Type", "Reference_Allele", 
        "Tumor_Seq_Allele1", "Tumor_Seq_Allele2")]
    anno.df2 = anno.df[, colnames(anno.df)[!colnames(anno.df) %in% 
        colnames(anno.df1)]]
    anno.df = cbind(anno.df1, anno.df2)
    if (!is.null(basename)) {
        write.table(anno.df, paste(basename, "maf", sep = "."), 
            quote = F, row.names = F, sep = "\t")
    }
    return(anno.df)
}

Hello,

I am having issues with mutations that are not SNPs. For example, although oncotator has no problem annotating this, when I include it as part of a larger tsv file and run your script or just run your script on a tsv file that only has this mutation, I get an error :
Error in rbind(deparse.level, ...) : numbers of columns of arguments do not match

This is the mutation I have been using as an example to test the script: 7 50367240 50367256 AAAGCCCCCCTGTAAGC AGTGGGG.

Thanks so much for this script!