Re: [crux-users] tide-index Fatal I/O error writing peptides to temp files

18 views
Skip to first unread message

William S Noble

unread,
Nov 7, 2019, 7:48:06 PM11/7/19
to Amy Zimmerman, crux-users
Hi Amy,

I am not able to reproduce this error. I suspect that you are filling up your temporary disk space. On linux systems, this is typically located at /tmp. You can run this shell command to see is your system is configured otherwise: echo ${TMPDIR:-/tmp}. On Windows it is usually C:\Users\<username>\AppData\Local\Temp, or you can run echo %TMP% or echo %TEMP% to see if your system is configured otherwise. Note that you can use the crux --temp-dir  parameter to specify any other location to store crux temporary files, though specifying a network drive can affect the performance.

However, you should keep in mind that with the parameter you are using, you are experiencing a combinatorial explosion of modified peptide sequences.  Just as an example, if you use your parameter settings on a database that contains only the very first, short sequence from your database (VSALKLQGGQHGGSAYNDVPKTLFFVPEKVQKQGLEK) then the resulting index contains 572,775 peptides. Searching with a very large database will almost certainly lead to a big loss of statistical power to identify spectra. At the very least, you might want to consider using a cascade search strategy, to first identify unmodified spectra and then sequentially search for more types of modifications.

Good luck, and please let us know if you have further questions.
Bill

On Thu, Oct 31, 2019 at 12:13 PM Amy Zimmerman <aezim...@uchicago.edu> wrote:
Hello,

I am trying to analyze peptide data from a 15N-incorporation experiment by running crux (crux-3.2.3-el7-x86_64) on our university's high performance compute cluster. However, when I try to run tide-index with modified parameters that allow for variable modifications (i.e., different levels of heavy nitrogen incorporated into each peptide), I have consistently been getting a fatal I/O error after the program has written ~944,000,000 peptides to temp files. The peptide database I'm trying to index has 682,415 entries in it. Memory does not seem to be an issue, but I'm not sure how to troubleshoot this. For comparison, I have been able to successfully run SEQUEST searches in ProteomeDiscoverer using the same peptide database and allowing for up to 8 variable modifications per peptide.

I've attached two of the parameter files that I've tried-- one allows for 4 variable modification, the other allows for 8 variable modifications per peptide (unless I am misunderstanding these parameters). The peptide database (fasta) that I'm trying to index should be accessible to members of this group through the following link:

Below is an example of the head and tail of the output I get when I've executed "crux tide-index" with the attached parameter files and database:

WARNING: Read parameter 'post-processing-qvality' from parameter file, but no such parameter exists.
INFO: Writing results to output directory '/project2/mlcoleman/project/hot_proteomics/crux-output/index_db3_max8/'.
INFO: Crux version: 3.2
INFO: Thu Oct 31 08:34:14 CDT 2019
INFO: Beginning tide-index.
INFO: Running tide-index...
INFO: Writing results to output directory 'db3_max8.index'.
INFO: Reading /home/aezimmerman/hot_proteomics/databases/db3_HOTBATSsrf_RNA.fasta and computing unmodified peptides...
INFO: Ignoring 105670 peptide sequences containing unrecognized characters.
INFO: Generated 7326227 targets, including duplicates.
INFO: Writing reverse-protein fasta and decoys...
INFO: Generated 7967782 decoys.
INFO: Generating 1 decoy per target
INFO: Reading proteins
DETAILED INFO: 15294009 peptides in heap
INFO: Wrote 100000 peptides
INFO: Wrote 200000 peptides
INFO: Wrote 300000 peptides
INFO: Wrote 400000 peptides
INFO: Wrote 500000 peptides
INFO: Wrote 600000 peptides
INFO: Wrote 700000 peptides
INFO: Wrote 800000 peptides
INFO: Wrote 900000 peptides
INFO: Wrote 1000000 peptides
...
INFO: Wrote 944420000 peptides to temp files
INFO: Wrote 944430000 peptides to temp files
INFO: Wrote 944440000 peptides to temp files
INFO: Wrote 944450000 peptides to temp files
INFO: Wrote 944460000 peptides to temp files
INFO: Wrote 944470000 peptides to temp files
INFO: Wrote 944480000 peptides to temp files
FATAL: I/O error writing modified peptide 

Any guidance would be much appreciated.

Best,
Amy

--
You received this message because you are subscribed to the Google Groups "crux-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crux-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/crux-users/93b0240d-47cf-4720-b3da-a9feb822dc93%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages