some documentation suggestions

[will-rowe]

Hi Alessia,

Firstly, thank you for releasing this pipeline - it's a good workflow and is working nicely for us.

I just had a couple of suggestions after having installed and run YAMP on a few different systems. I thought it might be useful to make a note here in case others were having the same difficulties.

• the --prefix option can't contain a #

  • I found this out as I was using seq data basenames as a prefix (e.g. --prefix 123_4#5) . It's not a big problem but caused a bit of headache as there was no error message (pipeline just terminated early)

• contam_15 contains a P char in sequencing_artifacts.fa in downloaded resources

  • BBMap (v 37.02) didn't like this and the pipeline would through an error.

• really minor again but the uniref path path in nextflow.config did not match the resources dir structure (when everything else did)

Thanks again (and sorry if I have missed these points anywhere in the wiki already).

Will

[alesssia]

Hi @will-rowe,

thanks for your feedback!

I've now added a Troubleshooting page to the wiki here, and also fixed the uniref path in the Folder layout example. Everything should look good now. Thanks a lot for pointing out the # problem!

Regarding the second point you raised. I've downloaded the sequencing_artifacts.fa from the BBmap webpage (it is one of its component) and used it with both BBmap v36.92 and v37.10. Can you send me more details on the error so I can investigate it?

Thanks a lot, Alessia

[will-rowe]

No worries! Here is how I produce the BBMap error. Although I made a mistake, it was v37.78 (most recent version in bioconda) that causes the error. Downgrading to v37.02 fixes it!

• downloaded YAMP_resources_20171128.tar.gz

• followed instructions to index and get the dir layout correct

• ran the example command:

  ../nextflow run YAMP.nf --reads1 ./data/ERR011089_1.fastq.gz --reads2 ./data/ERR011089_2.fastq.gz  --prefix Meta_HIT_ERR011089 --outdir ./data --mode complete

• receive this error:

  java.lang.Exception:
  An input file appears to be misformatted:
  The character with ASCII code 67 appeared where a base was expected.
  Sequence #15
  Sequence ID 'contam_15'
  Sequence: [65, 67, 65, 67, 71, 84, 67, 84, 71, 65, 65, 67, 84, 67, 80, 65, 71, 84, 67, 65, 67, 78, 71, 65, 84, 67, 71, 71, 65, 65, 71, 65, 71, 67]
  ACACGTCTGAACTCPAGTCACNGATCGGAAGAGC

Fixed with downgrading BBMap, or remove the P (or edit bbmap commands to fixjunk).

[alesssia]

Thanks for the very detailed report! I will look into this tomorrow. If I may ask, what are you working on?

[will-rowe]

You're welcome. I work at the STFC on a few different metagenomics projects; mainly with gut and soil samples. I'm also working on metagenomics tools for resistome profiling etc.

[alesssia]

Sounds cool!

if you are using YAMP to process soil samples, do you mind sharing some stats on computational requirements? I've run a few datasets to check feasibility, but I'd love to have some more information on real-life scenarios. Also, if you want to share some wisdom on how selecting the best processing parameters for processing non-human samples (we have been working only on gut so far) I may create a wiki page (full credit to you, obviously) that may be helpful to other users. Thanks!

[will-rowe]

Sure, happy to send some info on. I plan on doing a set of experiments in a few weeks time so will post the performance when I get it.

Not sure I have any wisdom re. best parameters but I will make some notes over the next few weeks and then see if they are any use to you!

[alesssia]

Regarding the open BBmap issue. I've added the problem to the troubleshooting page. Thanks for pointing it out and for accepting to share stats and wisdom!

I am also closing this issue!