Error running Trinity

Question

[Guilherme]

Hello, I'm new to bioinformatics and to linux command line so I'm having small issues troubleshooting in an assignment. Anyone know why I can't run Trinity here?

The tutorial I'm following tells me to use the following command:

 ~/bin/trinity/Trinity --seqType fq --max_memory 2G --CPU 2 --SS_lib_type RF --left rna_data/Sp_ds.left.fq.gz,rna_data/Sp_hs.left.fq.gz,rna_data/Sp_log.left.fq.gz,rna_data/Sp_plat.left.fq.gz --right rna_data/Sp_ds.right.fq.gz,rna_data/Sp_hs.right.fq.gz,rna_data/Sp_log.right.fq.gz,rna_data/Sp_plat.right.fq.gz --trimmomatic

And I get the following error message:

Error, cannot locate file: rna_data/Sp_ds.left.fq.gz at /home/user/bin/trinity/Trinity line 2550.
        main::create_full_path('ARRAY(0x1d15680)', 1) called at /home/user/bin/trinity/Trinity line 1207

Anyone care to explain how i can solve this? The files are in the rna_data directory just as they're supposed to be, I've checked

[Jack Tierney]

Where is the rna_data directory relative to where you are running Trinity? The error suggests it's an issue with paths. One option might be to use absolute paths instead of these relative ones

[Guilherme]

Ty very much! Absolute path worked After some trial and error i also figured out I was in /rna_seq/rna_data and I should have been running trinity from rna_seq

Answer 1

[brian.fristensky]

The first thing to try is running the Trinity command from within the rna_data directory, which means you would have to remove "rna_data/" from the command. Every program is different in how it handles file paths, and a quick review of the man pages doesn't give insight into what Trinity does. The other advantage is that you keep the command line simple, which makes de-bugging easier. I always run Trinity from within the directory containing the reads.

It's also a good organizational practice to use the --output <dirname> option to send all output to a separate directory.

Do the files in the rna_data have .fq.gz extensions, or did the previous step (eg. error trimming, error correction) generate uncompressed .fq files? If the files actually end in .fq, then Trinity would give the above error when looking for files ending in .fq.gz. The other thing the help pages don't address is whether or not Trinity accepts gzipped read files.

Finally, you might want to have a look at the BIRCH system, which automates many genomics/transcriptomics tasks through an easy to use graphical interface called BioLegato. The Transcriptome Assembly tutorial shows you can point-and-click your way through assembling a transcriptome, and not have to worry about the maddening array of filename and file format issues that come with the territory.

[Guilherme]

Thank you for the comment, as I said I'm still a newbie and this is a huge help!