-1
$\begingroup$

I have ~130 gene names of different human proteins. I'm looking for a convenient and systematic way to locate each gene's location in the genome, extract its CDS sequence and find whether there are SNPs in the CDS (in any context) and what's their position. Moreover, I've chosen for each of those proteins, several amino acids for whom I'm trying to find if the SNP fall in their exact coding nucleotide region (the amino acid triad of nucleotides).

Alternatively, I'm looking for an algorithm that would take the gene name out of my list, find its location in the genome and look for SNPs in the exons - if it would found any SNPs, then it should translate the exon to an amino acid sequence and find whether one of my pre-defined amino acids from each protein is translated by the nucleotide site (triad of nucleotides) with the SNP.

I'm using R, and I thought to use web scraping (should I use the rvest package?) but I'm not quite familiar with the technique. Is it possible? What websites should I use? Genome Browser? EMBOSS Transeq? Other websites?

Thank you

$\endgroup$
2
  • 1
    $\begingroup$ Why do you need to do web scraping instead of using biomaRt and the packages of Bioconductor to find and retrieve the information? Did you look up how do other papers obtain the sequence and positions? I think your attempt parsing webs is full of problems if not unfeasible. $\endgroup$
    – llrs
    Commented Feb 4, 2018 at 8:44
  • 1
    $\begingroup$ "convenient and systematic way to map genes..." we can only dream. $\endgroup$
    – zx8754
    Commented Feb 5, 2018 at 9:06

2 Answers 2

1
$\begingroup$

You can search databases like Cosmic(Cancer) with the gene names to get frequently occurring mutations in cancer. I'm not sure about any specific tool/pipeline available for this exact scenario.

As an alternative approach literature mining for gene associated mutation discovery, you can use text mining with machine learning to automate the whole process.

With no specific context in mind, you will end with several combinations of mutation per gene. Then for each variation translate the DNA sequence and match the result with reference peptide sequence you have.

$\endgroup$
1
$\begingroup$

I don't have a complete answer to your question, but as a primer I can tell you that you should look into the Bioconductor packages. The reason is that biological sequence data are deposited in online databases (like GenBank, and others), and exposed by these databases in such a way that you precisely don't need web scrapping techniques to retrieve the data you're interested in.

rvest is a great package for web scrapping, but definitely not the appropriate tool to solve your problem.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.