FAIL: meme-chip1 (differences in meme-chip.html)

[Teshome Mulugeta]

Hi,

Installing meme_4.12.0 failed.

last few lines from tests/scripts/meme-chip.log

meme-chip_html_to_tsv ran successfully in 0.203019 seconds

-------------------------------------------------------------------------------

A difference was detected between the reference and tested json files.

Got data:programs@6:outputs but expected data:programs@6:messages_file.

FAILED!

Best,

Teshome

[CharlesEGrant]

This indicates that meme-chip on your system is generating different output files then expected for the test case. This could be benign, or it could indicate a genuine issue. In the directory tests/results/meme-chip1 you should find a file meme-chip.html. If you could forward that file to us it would help troubleshoot the issue. It would also be helpful to run 'ls' in that directory and forward us the results. 

Of course this is only relevant if you are planning to run meme-chip. If you just want to run fasta-grap, you can go ahed with 'make install'.

[Teshome Mulugeta]

I couldn't find a way to attach the requested file here but i replied to the email i received.

[CharlesEGrant]

When you are editing a post there is a paper-clip icon just above the text editing toolbar, next to a link labeled "Attach a file".

[Teshome Mulugeta]

I am sorry, I am not able to see the paperclip. Screenshot is attached if it is visible.

[Teshome Mulugeta]

I tried to debug it myself by looking at the difference between the two files and the eye-catching difference is the version and revision. I downloaded 4.12.0 but the test html came with older version and revision ids. Is this correct?

$> diff meme-chip/meme-chip.html results/meme-chip1/meme-chip.html

10,12c10,12

<         "version": "4.11.3",

<         "revision": "50202ae415e2a326ff31b6ecfc65dd4910dca68c",

<         "release": "Fri Feb 19 13:23:06 2016 -0800",

---

>         "version": "4.12.0",

>         "revision": "aacc9c3eb8c7b460b344d5c8aa1458097dae704d",

>         "release": "Tue Jun 27 16:22:50 2017 -0700",

24c24

Best,

Teshome

[CharlesEGrant]

No, the testing script knows to ignore items like the version number, dates, or timing information. Using the copy of meme-chip.html you mailed me I was able to spot the issue:

            "name": "meme",

            "cmd": "meme results/meme-chip1/seqs-centered -oc results/meme-chip1/meme_out -mod zoops -nmotifs 3 -minw 6 -maxw 30 -bfile results/meme-chip1/background -dna -revcomp -nostatus",

            "status": 0,

            "oot": false,

            "time": 79.131631,

            "messages_file": "meme_msgs.txt",

            "outputs": [

This is confirmed by the listing you provided for the output directory. For some reason when MEME is run on your system it generates a message file, meme_msgs.txt, that is not normally generated when MEME ChIP runs. Take a look at the contents of meme_msgs.txt. If the messages is innocuous, then the test failure can be ignored.

[CharlesEGrant]

Hmm. That's odd. I checked using current versions of  Safari, Chrome,and Firefox, and each has an edit menu containing an "attach file" item, as in this screen shot:

You'll have to check with your local computer support about this.

[Teshome Mulugeta]

Thank you for pointing meme_msgs.txt. The problem was that it needed path to the correct ghostscript provided --with-gs=

The error message gone when we provided that and we used python 2.7 instead of 3.5.

./configure --prefix=/local/genome/packages/meme/4.12.0.test --with-url=http://meme-suite.org --enable-debug --with-perl=/local/genome/packages/perl/5.18.2/bin/perl --with-python=/local/genome/packages/python2/2.7.6/bin/python --with-mpidir=/local/genome/packages/openmpi/1.10.1/ --with-gs=/local/genome/packages/ghostpdl/9.20/bin/gs

Best,

Teshome

[Linda Molla]

Hello, 

I also get the same error:

FAIL: meme-chip1 (differences in meme-chip.html)

python --version

Python 2.7.12

I can also send you the output of the meme-chip1 but I also can not attach them here  /tests/results/meme-chip1

The content of meme_msgs.txt : 

[[52211,1],0]: A high-performance Open MPI point-to-point messaging module

was unable to find any relevant network interfaces:

Module: OpenFabrics (openib)

  Host: ip-172-31-17-154

Another transport will be used instead, although this may result in

lower performance.

[CharlesEGrant]

This message is harmless and can be ignored. There is something sub-optimal about the configuration of OpenMPI on you local machine which produces that warning message, but it is only a warning message. It's not something we get on our development machines, which is why its flagged as a test failure.