When Bioinformatics APIs break

NCBI assembly FTP

I wrote a Node.js module for accessing the NCBI API. Yesterday tests passed, today they don’t.

Yesterday, doing for example:

bionode-ncbi search assembly Guillardia theta

would give this. Now it’s returning this instead.

The difference between the two results is the following:

< "FtpPath" : {
< "type" : "GenBank",
< "_" : "ftp://ftp.ncbi.nlm.nih.gov/genbank/genomes/Eukaryotes/protozoa/Guillardia_theta/Guith1/"
< }
---
> "FtpPath" : [
> {
> "type" : "GenBank",
> "_" : "ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCA_000315625.1_Guith1"
> },
> {
> "type" : "RefSeq",
> "_" : "ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCF_000315625.1_Guith1"
> }
> ]

Which seems to indicate that the location for the genome assembly changed. So I just need to update my tests right? But wait, the new location doesn’t exist while the old one is still available. >_<

So now I don’t have a way to automatically get the genome assemblies I need for my work. Let’s hope this is just an unannounced change followed by a slow migration, and soon the files will show at the new location.

The FTP links on their website are also broken (panel on the right side). If you get an “Invalid BlobId” error with the previous link, try copy/pasting it to a incognito mode tab in Chrome… :-/

In the meantime, I guess I’ll try to get the sequences from EBI