Salmon index error

[Devinder Kaur]

Hello,

I was running salmon for RNA quantification.
Using "salmon index", the index built successfully. But I faced following problem when run the "salmon quant" command:
Error: The index version file index/versionInfo.json doesn't seem to exist.  Please try re-building the salmon index.

Version I am using: ### salmon (mapping-based) v0.8.2

[Rob]

Hi Devinder,

Can you please write the commands you used? Also, it would be very useful to see the directory listing and file sizes for the directory where you told salmon to write the index. It appears it can't find a file it expects; suggesting the file is missing or otherwise unaccessble.

Thanks,
Rob

[Devinder Kaur]

Here is the command for Index building:

devinder@devinder-desktop:~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64/bin$ ./salmon index -t ~/Desktop/HM1-transcript-fmt.fasta -i ~/Documents/softwaret
ools/Salmon-0.8.2_linux_x86_64/HM1run/HM1-transcript_index
Version Info: This is the most recent version of Salmon.
index ["/home/devinder/Documents/softwaretools/Salmon-0.8.2_linux_x86_64/HM1run/HM1-transcript_index"] did not previously exist  . . . creating it
[2017-05-08 16:33:13.237] [jLog] [info] building index
RapMap Indexer

[Step 1 of 4] : counting k-mers
Elapsed time: 0.757082s

Replaced 9409 non-ATCG nucleotides
Clipped poly-A tails from 0 transcripts
Building rank-select dictionary and saving to disk done
Elapsed time: 0.0023778s
Writing sequence data to file . . . done
Elapsed time: 0.0158964s
[info] Building 32-bit suffix array (length of generalized text is 10471711)
Building suffix array . . . success
saving to disk . . . done
Elapsed time: 0.0557196s
done
Elapsed time: 1.59784s
processed 10000000 positions
khash had 9192283 keys
saving hash to disk . . . done
Elapsed time: 0.893572s
[2017-05-08 16:33:28.478] [jLog] [info] done building index

In the attached picture you can see the folder of index with details of files generated.

[Rob]

Hrmm --- everything looks to be in order here.  You can see from the screenshot that the "versionInfo.json" file seems to exist (can you list its contents?).  What is the command you use to run Salmon?  From this index alone, I don't see any red flags.

Best,

Rob

[Devinder Kaur]

Hi Rob,

To bring into your notice, I simply download the appropriate binary (Salmon-0.8.2_linux_x86_64) and uncompress it. The salmon located in side the bin, so I simply run it using command :

devinder@devinder-desktop:~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64/bin$ ./salmon -h
Salmon v0.8.2

Usage:  salmon -h|--help or
        salmon -v|--version or
        salmon -c|--cite or
        salmon [--no-version-check] <COMMAND> [-h | options]

Commands:
     cite  Show salmon citation information
     index Create a salmon index
     quant Quantify a sample
devinder@devinder-desktop:~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64/bin$

In the installation guide it was mentioned to go for "make, make install" but when i did it, it was showing :

devinder@devinder-desktop:~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64$ make
make: *** No targets specified and no makefile found.  Stop.
devinder@devinder-desktop:~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64$ make install
make: *** No rule to make target 'install'.  Stop.
devinder@devinder-desktop:~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64$ make test
make: *** No rule to make target 'test'.  Stop.

then I tried configuring it, but still it was not done.

devinder@devinder-desktop:~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64$ ./configure
bash: ./configure: No such file or directory
devinder@devinder-desktop:~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64$ ./configure
bash: ./configure: No such file or directory

[Rob]

Hi Devinder,

  If you downloaded the pre-compiled binary, there is no need to make or build the software (it's already compiled).  This is why you're not able to run make / make install.  Those instructions are only for if you wish to build salmon from source.  With the pre-compiled binary, you simply run it as you showed in your first example.  Given the index you have, you'd have a command something like:

$~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64/bin/salmon quant -i ~/Documents/softwaret

ools/Salmon-0.8.2_linux_x86_64/HM1run/HM1-transcript_index -l A -1 location_to_reads_1.fq.gz -2 location_to_reads_2.fq.gz -o ~/Documents/softwaret

ools/Salmon-0.8.2_linux_x86_64/HM1run/HM1-transcript_quant -p 8

or something like this depending on the type of sample you're trying to quantify and the location of your reads.

Best,

Rob

[Devinder Kaur]

Hi Rob,

salmon is working now.
It was fault at my side. I was using following command: where the mistake text is in blue.

devinder@devinder-desktop:~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64/bin$ ./salmon quant -i index ~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64/HM1run/HM1-transcript_index -l IU -1 ~/Documents/strandspecific-Illumina-HM1/read-files/HM1-NORMAL_R1.fastq -2 ~/Documents/strandspecific-Illumina-HM1/read-files/HM1-NORMAL_R2.fastq -o ~/Documents/softwaretools/Salmon-0.8.2_linux_x86_64/HM1run/test-out

THANK YOU so much for clearing the matter.

ONE Question I want to ask, you suggest to use thread -p 8, is it most widely and acceptably used or we can make it flexible. What would be the strong justification of using it at -p 8 ?

[Paul Dellorusso]

Hi Rob,

I'm running into a similar issue as Devinder was and having trouble figuring it out. Was hoping you could help. Get the same error message saying my index/versionInfo.json does not seem to exist (screenshot attached). I have single end reads spread across 5 fastq files per sample. Just trying to run one sample as a test before looping it. 

I installed the latest version per our discussion directly to my bin and then removed salmon from the conda install. I've attached a screenshot of the indexing process (it boots me out at the end but I cant tell from the outputs if it was cut short, last thing it notes is writing sequence data to the file). I've also attached a screenshot of the generated index folder contents.

I tried indexing a transcripts FASTA file from Gencode as well and am running into the same issue, so dont think its the specific reference. The reference transcript files are 54 and 56mb respectively, so no issues downloading and transferring to our cluster. 

 

Any thoughts are much appreciated! Thanks!

Paul