Locating BLAST data needed by netsurfp

Question

[lefthandgergo]

Hi,

I am trying to install and configure netsurfp (https://services.healthtech.dtu.dk/service.php?NetSurfP-2.0) for one of our projects. It needs the location of BLAST data on my system (example given in the script: "/usr/cbs/bio/src/blast-2.2.18/data/"). I am using Ubuntu 20.04, I have installed both ncbi-blast+ and ncbi-blast+-legacy packages, but could not find such directories anywhere. I also downloaded the official BLAST packages from the NCBI FTP server (http://mirror.ufs.ac.za/datasets/ncbi/blast/executables/blast+/2.2.18/), but I could not find any data directory inside them. Do you have any idea where to look for them?

Thanks,

Gergő

[lieven.sterck]

without knowing the specific details I assume it wants the location where you installed your blast DBs (can be a separate folder than where the binaries of blast are located).

if you correctly installed the blast tools, you can for instance run which blastp that will give you the location of the blast binary files

on a side note: the blast version 2.2.18 is years old (dates from 2013), if in any way possible download a more recent version of blast.

Answer 1

[Mensur Dlakic]

As @GenoMax told you, this refers to legacy BLAST versions. When you unpack the packages, they contain a data directory which contains substitution matrices and similar things. I think that is the directory they refer to in NetSurfP installation instructions.

Latest legacy BLAST is probably what you want:

https://ftp.ncbi.nlm.nih.gov/blast/executables/legacy.NOTSUPPORTED/2.2.26/

[lefthandgergo]

Thank you for the help. Indeed this was the missing directory I was looking for!

Answer 2

[GenoMax]

What you are looking at is called legacy blast due to the fact that this software has not been supported for a long while (v.2.2.18 is probably from 2008). Said `data` directory likely originally contained blast database indexes. I don't recollect if old blast package used v.4 format indexes. NCBI still keeps an old version of v4 blast indexes from early 2020 available here. You will need to create those from relevant fasta files, if these don't work.

Edit: Looks like you can find functionality of NetSurfP in hh-suite (LINK) which is currently maintained. They also have a nice user guide available.

[lefthandgergo]

Thanks for the answer, I will try out your suggestions. In fact I would like to use BepiPred (http://www.cbs.dtu.dk/services/BepiPred/), which requires netsurfp.