Running splign just like NCBI

Question

[Reece]

NCBI exposes (e.g., via eutils) genomic coordinates for current transcripts on the current assembly only. I would like coordinates for older transcript versions on newer assemblies and for newer transcripts on older assemblies, and be reasonably confident that I'm mimicking NCBI behavior as closely as possible.

Has anyone out there reproduced NCBI's invocation of splign? If so, how?

To be clear, I have splign and friends running. I'm really asking about the command line invocation and data prep.

 

[dukecomeback]

Hi, I was wondering if you are keep using splign for transcript mapping. I also got the output from splign and could you kindly give me some suggestion on how to parse and filter it? Thank you very much.

Answer 1

[Reece]

Based on conversations with several people at NCBI ~5 years ago, I determined that the data prep and post-processing steps employed by NCBI (and unreleased) would make it essentially impossible to reproduce their results.

I resorted to making snapshots of NCBI's gff files (ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/alignments/), which contain the resulting alignments for everything that I care about.

[dukecomeback]

Hi Reece, Thank you very much for your suggestions!