Blast+ in RAVEN

[parsanayyara]

Hello,

I noticed that RAVEN uses BLAST+ to run getBlast() for reconstructions, I followed the BLAST+ installation as noted on the website and install it on the RAVEN folder, but everytime I tried running checkInstallation on RAVEN, it kept giving this error. Can anyone give suggestions on how to properly get Blast+ installed for RAVEN?

[mihai-sysbio]

Hi @parsanayyara! Could you specify what operating system and what version of Matlab you are using?

[parsanayyara]

MATLAB R2022b in 64-bit Windows

[mihai-sysbio]

The executables for BLAST+ (as packaged with RAVEN) haven't been updated in two years https://github.com/SysBioChalmers/RAVEN/tree/main/software/blast%2B. We are distributing these, so there is no need to install BLAST+ separately. For testing purposes, could I ask you to fetch an older release v2.7.12 and see how that one behaves?

[parsanayyara]

I will try and do that. On another note though, suppose I manage to get it RAVEN up and running, would it be possible to do a reconstruction only based on MetaCyc and can the draft reconstruction generated by RAVEN be exported into excel or is it already in COBRA-compatible SBML?

[mihai-sysbio]

Instead of risking to not be giving the fully accurate answers here, I'm pinging @edkerk.
In general, I wouldn't worry about the export. To my knowledge, the SBML is COBRA-compatible, and there is Excel export as well.

[edkerk]

Regarding your MetaCyc question, you can make a reconstruction based only on MetaCyc but I would not recommend this as MetaCyc-based reconstruction tend to contain many false positive hits. Instead, MetaCyc-based reconstruction is best used as complementary approach, where the reactions are checked and curated before added to another (draft) model.

Finally, RAVEN exports valid SBML L3V1 FBCv2 files, which are COBRA-compatible. I strongly discourage the use of Excel as intermediate file format. Excel can be great if you quickly want to look up some aspects of the model (and are a proficient Excel-user), but should not be used to make changes to the model.

[edkerk]

I cannot replicate the non-functioning BLASTP executable, you can perhaps run blastPlusTests to get more information.

[dimwas]

I can confirm the problem exists; strangely enough, it occurs on my workplace PC, but not on my laptop and home PC. On all of them I use 64-bit Windows and MATLAB R2023a. I ckecked two RAVEN releases (2.8.7 and 2.9.2), the result is the same. I tried to run blastPlusTests, the results are not very informative:

Name: 'blastPlusTests/testBlastPlus'
         ProcedureName: 'testBlastPlus'
             TestClass: [0×0 string]
            BaseFolder: 'D:\RAVEN\testing\unit_tests'
      Parameterization: [0×0 matlab.unittest.parameters.EmptyParameter]
    SharedTestFixtures: [1×0 matlab.unittest.fixtures.EmptyFixture]
                  Tags: {1×0 cell}

Tests Include:
    0 Parameterizations, 0 Shared Test Fixture Classes, 0 Tags.

I also tried

1. to reinstall RAVEN;
2. to use blast+ version from NCBI website;
3. to copy blast+ folder from a PC where it works.

It didn't work. Could you maybe suggest some steps to resolve the issue? @parsanayyara, were you able to find a solution?
Edit: I tried the release v.2.7.12 suggested in the answer - the problem persists.

[dimwas]

Some progress, but the problem in New DB title:

message =

    '
     
     Building a new DB, current time: 06/26/2024 16:08:41
     New DB name:   c:\tmp\tp3b366439_496f_4d3e_b56f_9d1fdf6b9724\tmpDB
     New DB title:  C:\Users\firstname lastname\AppData\Local\Temp\tpeb35645a_29c5_4b53_817e_704da2eb538e
     Sequence type: Protein
     Keep MBits: T
     Maximum file size: 1000000000B
     BLAST options error: File C:\Users\firstname does not exist
     '

[edkerk]

Make the same change on line 72 of checkFileExistence.m

[dimwas]

Aaand now it works :) Great, thank you; will this new tmp folder be deleted automatically?

[edkerk]

I don't think so, will look into this when properly implementing it in RAVEN.

[edkerk]

Solved in #551