Species Not Present in the OMA Database

[HHaoyi]

Hello, amazing work!

I have a question regarding the annotation of non-model species that are not present in the OMA Database. To obtain orthologous protein evidence, I first ran OMA Standalone using protein sequences with default parameters, and I confirmed that the EstimatedSpeciesTree.nwk aligns with my expectations. This yielded an OrthologousGroups.orthoxml file. However, the Extract Consensus Annotation step failed with the following error:

INFO:pyham.ham:Parse Orthoxml: 21740 top level hogs and 86018 extant genes extract.
INFO:pyham.ham:Set up Ham analysis: ready to go with 21740 hogs founded within 2 species.
Traceback (most recent call last):
  File "/strage_151/home_187/wuyh/soft/OMAnnotator/OMAnnotation/OMAnnotation.py", line 494, in <module>
    args.func(args)
  File "/strage_151/home_187/wuyh/soft/OMAnnotator/OMAnnotation/OMAnnotation.py", line 92, in extract_consensus
    cons_fasta, cons_gff = select_consensus_sequence(protids, gff_corr, fasta_corr)
  File "/strage_151/home_187/wuyh/soft/OMAnnotator/OMAnnotation/OMAnnotation.py", line 261, in select_consensus_sequence
    hog_record_list = [corr_fasta_map["_".join([hid.split(' ')[0], hif])] for hid, hif in hog_prot_id]
  File "/strage_151/home_187/wuyh/soft/OMAnnotator/OMAnnotation/OMAnnotation.py", line 261, in <listcomp>
    hog_record_list = [corr_fasta_map["_".join([hid.split(' ')[0], hif])] for hid, hif in hog_prot_id]
KeyError: 'evm.model.Hic_asm_11.2212_china_protein'

My question is: Could this error be due to the fact that I ran OMA Standalone using protein sequences, but for OMAnnotator I used genome files and annotation files? Or is it because I did not run the "Preparing the data" step?

In other words, when working with species that are not present in the OMA Database, what is the recommended workflow for using OMAnnotator? Do you have any suggestions for an improved process?

Any help or advice would be greatly appreciated!

Best regards

[andreaswallberg]

Hi!

I might have a related issue, although a different error message. In my case, I get many repeats of this when running OMA Standalone:

Error, (in _check_species_tree_param) 
SpeciesTree does not contain the same leaves as sp ..(1054).. 7','cpac8'}

It does not matter if I remove the FASTA and GFF datasets associated with the species in the message (cpac8) - then I get the same error on the next species. To me, it seems like the information for how to proceed under "Example step 2: preparing the data" is incomplete.

Any hints are much appreciated!
Cheers!

[YanNevers]

Dear @HHaoyi ,

I apologize for the delay in answering.

I think this may indeed be due to not running the prepare data step at first, which is normally used by OMAnnotator to extract the protein sequences.
The recommended workflow is indeed to:

1. Run the prepare_data step of OMAnnotator
2. Run OMAStandalone once the data have been prepared
3. Extract the consensus.

The reason why we recommend using prepare_data even when one have a FASTA file is that OMAnnotator needs clean way to be able to map the GFF information to the protein FASTA file, which would be complicated when the FASTA file is generated by any other tool.

In your case, it may still be possible to conclude the OMAnnotator process without redoing all previous steps.
You can:

1. Check he the protein identifier in your FASTA file is in the same model as the start of the identifier generating the key error.
2. If it is the case, edit the headers of the FASTA file to be in this form ">{protein_id] origin={file_prefix}. Keeping the protein_id as is in the original file, with file prefix being the filename string, excluding the extension.

Let me know if it works, or if you need more detail. You can also contact me via email if you wish to work out a solution directly.

Best wishes,
Yannis

[YanNevers]

Dear @andreaswallberg ,

Could you send me the species tree string that is in your 'parameters.drw' file?
If you did add other species that could be used homology evidence than the one originally present in OMA to your dataset, it is needed to edit the species tree to include those as well.

I realize it is not explicitly stated in the README, since our example does not include adding other evidence data. If this is indeed the issue, I will edit the README to make this clear.

If this is something else, I can help figuring it out and improve the documentation accordingly once it is done.

Thank you,
Best regards,
Yannis