biocore / deblur Goto Github PK

See here

positive (gg 88% rep set) and negative (phix+adapter) fasta files are a part of the git
need deblurring to access them by default based on the mode (-n)
also need to test if not indexed, index them (1st time or as part of the install) and then use the indexed versions by default to save re-indexing each run...

• example install via pip or conda
• basic command shown should reflect the click command structure
• add change log

Having the 100 character names makes it very difficult to read the biom tables when converted into pandas. We really should be using some sort of unique hash generated from the sequence as OTU ids, for the sake of readability.

need to warn the user if he deblurs using trim longer than the run read length (i.e. -t 150 and the sequencing is 100bp long)

Strongly recommend removing biom 1.x format support and requiring h5py.

mafft failes (only 1 sequence to align)
so a samples containing only 1000 identical reads will fail in the mafft stage

Don't throw away the read error reads, but instead add back to the original sequence

We are only correcting by reducing the neighbor frequency, unsure how this affects if we increase back the frequency of the current sequence.

otherwise user needs to supply/compile the 88_otus.fasta

every X% (default 5) of samples (and also indexing of GG and splitting...)
have a flag for X, 0 if supress

just need to change the one import for stringio to include if version

Looking at some of the taxonomy metadata, the last few characters are cropped off. Here is an example of such a taxonomy string.

['k__Bacteria',  'p__Proteobacteria',  'c__Alphaproteobacteria',  'o__Sphingomonadales', 'f__Sphingomonadaceae']

Notice how the last few items are cropped off (i.e. genus and species). This makes it more difficult to load the taxonomies into a pandas dataframe.

It would be great if some padding could be done in the post processing.

install the artifacts.fa and 88_otus.fasta files

I.e. even if are low level sequences do the math.

Its not clear exactly what dependencies are required to run this.

After talking to @josenavas, I realized that MAFFT and SortMeRNA are required.
What other dependencies are required? The documentation will need to be updated with this information, so we can turn this into a conda recipe.

See original discussion here

maybe -j or -n instead of -O?
(but -O is in qiime?)

I'm getting a weird error when I tried to run the deblur workflow.

I'm running the following command

deblur workflow \
  --seqs-fp 648_seqs.fasta \
  --output-dir deblurred \
  --ref-fp /home/mortonjt/deblur_db/artifacts.fa \
  -n -w -d 1,0.06,0.02,0.02,0.01,0.005,0.005,0.005,0.001,0.001,0.001,0.0005 \
  -t 150 -O 32 \
  --log-level 2 \
  --log-file log.newmonkies.neg \
  --min-reads 25

And I get the following error

discarding /home/mortonjt/miniconda/bin from PATH
prepending /home/mortonjt/miniconda/envs/deblur/bin to PATH
Traceback (most recent call last):
  File "/home/mortonjt/miniconda/envs/deblur/bin/deblur", line 572, in <module>
    deblur_cmds()
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 716, in __call__
    return self.main(*args, **kwargs)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 696, in main
    rv = self.invoke(ctx)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 1060, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 889, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 534, in invoke
    return callback(*args, **kwargs)
  File "/home/mortonjt/miniconda/envs/deblur/bin/deblur", line 543, in workflow
    parallel_deblur(input_file_list, sys.argv, ref_db_fp, jobs_to_start)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/deblur/parallel_deblur.py", line 130, in parallel_deblur
    es))
RuntimeError: stdout: 
stderr: Traceback (most recent call last):
  File "/home/mortonjt/miniconda/envs/deblur/bin/deblur", line 572, in <module>
    deblur_cmds()
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 716, in __call__
    return self.main(*args, **kwargs)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 696, in main
    rv = self.invoke(ctx)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 1060, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 889, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/click/core.py", line 534, in invoke
    return callback(*args, **kwargs)
  File "/home/mortonjt/miniconda/envs/deblur/bin/deblur", line 535, in workflow
    delim=delim)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/deblur/workflow.py", line 570, in launch_workflow
    min_size=min_size, threads=threads_per_sample)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/deblur/workflow.py", line 82, in dereplicate_seqs
    sout, serr, res = _system_call(params)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/site-packages/deblur/workflow.py", line 679, in _system_call
    stderr=subprocess.PIPE)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/subprocess.py", line 711, in __init__
    errread, errwrite)
  File "/home/mortonjt/miniconda/envs/deblur/lib/python2.7/subprocess.py", line 1343, in _execute_child
    raise child_exception
OSError: [Errno 2] No such file or directory

exit: 1

The last few steps in log.newmonkies.neg was as follows

FO(139964386215680)2016-08-09 13:19:46,210:launch_workflow for file /home/mortonjt/Documents/potato_16S/data/illumina_reads/deblurred/split/1706.F54.7.prev.fasta
INFO(140568600663808)2016-08-09 13:19:46,315:dereplicate seqs file /home/mortonjt/Documents/potato_16S/data/illumina_reads/deblurred/deblur_working_dir/1706.F51.0.prev.fasta.trim
INFO(139964386215680)2016-08-09 13:19:46,732:dereplicate seqs file /home/mortonjt/Documents/potato_16S/data/illumina_reads/deblurred/deblur_working_dir/1706.F54.7.prev.fasta.trim
INFO(140204884723456)2016-08-09 13:19:46,739:dereplicate seqs file /home/mortonjt/Documents/potato_16S/data/illumina_reads/deblurred/deblur_working_dir/1706.F6.2.prev.fasta.trim

I'm not exactly sure what is going on. Any ideas about how I can start debugging this? Thanks!

The deblur function computes the hamming distance between the sequences twice. Check if caching those values is worth; i.e. test the trade-off between memory usage and running time.

need to set:

• read length?
• error profile
• positive or negative filtering
• fasta file for negative filtering
• indexed 88_otus.fasta or the 88_otus.fasta for positive filtering

following @ekopylova response, need to use sortmerna v2.1 and update the cli call inside workflow

why?

the people want taxonomy in their biom table
we can:

1. supply an additional script (using qiime rdp/other option?) to add taxonomy
2. add this script as an optional step in the pipeline (and have it as optional package dependency)

Yay, i'm writing an issue on github :)

1. When running the HMMER (or the phix detection) artifact filtering, we should output a summary of how many sequences were filtered out
2. Also add a flag/option to get a file with all the filtered out sequences (to see who they are...)

Depends on:

• #9
• #10
• #11
• #12
• #13
• #14
• #15

if -t not specified, automatically detect the read length (test XX first/random reads, take 80% percentile, if too varied show an error)

to enable log file analysis

Right now, this pipeline only takes in fasta files. Is there any interest in reading from fastq instead?

need only parse_fasta, remove_files
can copy instead

...and make sure to update classifications in setup.py

Got this from a user:
ANOTHER ERROR ( ONLY solution seems to be to divide the file up into pieces)

(deblurenv)MacQIIME dhcp238­043:~ $ deblur workflow ­­seqs­fp

/Volumes/Samsung_T1/MadaRaw/Mada_AS13JF14ND14_Splitlib_r3n0p.90q5/Mada_AS13JF14ND14_Splitlib_r3n0p.90q5_seqs250_254new.fasta

­­output­dir Madajul16_deblurtest ­O 3

Traceback (most recent call last):

File "/macqiime/anaconda/envs/deblurenv/bin/deblur", line 621, in

deblur_cmds()

File "/macqiime/anaconda/envs/deblurenv/lib/python3.5/site­packages/click/core.py", line 716, in call

return self.main(_args, *_kwargs)

File "/macqiime/anaconda/envs/deblurenv/lib/python3.5/site­packages/click/core.py", line 696, in main

rv = self.invoke(ctx)

File "/macqiime/anaconda/envs/deblurenv/lib/python3.5/site­packages/click/core.py", line 1060, in invoke

return _process_result(sub_ctx.command.invoke(sub_ctx))

File "/macqiime/anaconda/envs/deblurenv/lib/python3.5/site­packages/click/core.py", line 889, in invoke

return ctx.invoke(self.callback, **ctx.params)

File "/macqiime/anaconda/envs/deblurenv/lib/python3.5/site­packages/click/core.py", line 534, in invoke

return callback(_args, *_kwargs)

File "/macqiime/anaconda/envs/deblurenv/bin/deblur", line 561, in workflow

out_dir_split)

File "/macqiime/anaconda/envs/deblurenv/lib/python3.5/site­packages/deblur/workflow.py", line 445, in

split_sequence_file_on_sample_ids_to_files

outputs[sample] = open(join(outdir, sample + '.fasta'), 'w')

OSError: [Errno 24] Too many open files: '/Users/venceslab/Madajul16_deblurtest/split/J8.3.fasta'

I performed a pip install of deblur (pip install deblur) today and received the following traceback from a test run:

Traceback (most recent call last):
  File "/home/mcdonadt/miniconda3/envs/deblur/bin/deblur", line 621, in <module>
    deblur_cmds()
  File "/home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/site-packages/click/core.py", line 716, in __call__
    return self.main(*args, **kwargs)
  File "/home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/site-packages/click/core.py", line 696, in main
    rv = self.invoke(ctx)
  File "/home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/site-packages/click/core.py", line 1060, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/site-packages/click/core.py", line 889, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/site-packages/click/core.py", line 534, in invoke
    return callback(*args, **kwargs)
  File "/home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/contextlib.py", line 77, in __exit__
    self.gen.throw(type, value, traceback)
  File "/home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/site-packages/click/core.py", line 86, in augment_usage_errors
    yield
  File "/home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/site-packages/click/core.py", line 534, in invoke
    return callback(*args, **kwargs)
  File "/home/mcdonadt/miniconda3/envs/deblur/bin/deblur", line 570, in workflow
    working_dir=working_dir)
  File "/home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/site-packages/deblur/workflow.py", line 199, in build_index_sortmerna
    raise RuntimeError('Cannot index database file %s' % db)
RuntimeError: Cannot index database file /home/mcdonadt/miniconda3/envs/deblur/lib/python3.5/site-packages/deblur/support_files/artifacts.fa

It looks like the default values in the script are not the optimal ones, and/or the recommended ones. The default values should be what we think are the best parameters for running the commands.

cc @antgonza @amnona

sortmerna does not run on barnacle when obtained from biocore
it needs glibc 2.16
one workaround is to change to install instructions to use bioconda instead

need to make sure sample with 0 reads after singleton removal/msa (so 1 sequence before) doesn't fail

allow multiple demux fasta files as comma separated list

so we can have a non-default pos database for the post-deblur positive filtering