msarrias / exonize Goto Github PK

I think clarity would be improved if Blast-related code was refactored into a separate file and class. The Exonize object is quite heavy with many attributes and methods. If blast code was moved to a separate class, it could be instantiated in exonize.py and all its gory details would be "hidden" there, before being shared as an attribute to Exonize. Ideally, the blast abstraction would be so good that Exonize does not even need to know how sequences are compared, but that is not a priority.

review rules for classifying truncations.

clean up the counting of full duplications.

This relates to the functions insert_fragments_calls() in sql_utils.py and insert_fragments_table in exonize_handler.py.

• I was confused by the function insert_fragments_calls(). It returns two strings used to perform two SQL execute commands. I think it would make more sense to have functions for the two executes.
• insert_fragments_table has an internal function which in turn has an internal function. That is too many levels. The strand analysis could be done somewhere else, either when the blast results are collected in the first place or right before inserting into the DB tables.
• get_gene_tuple() returns a tuple with 6 elements. Element 3 is within parenthesis. Is there a reason for that? I printed the results and cannot see that it is needed.

After installing exonize following the instructions in the readme file I got the following warning:

Did you encounter the same issue?

/opt/homebrew/lib/python3.11/site-packages/Bio/init.py:138: BiopythonWarning: You may be importing Biopython from inside the source tree. This is bad practice and might lead to downstream issues. In particular, you might encounter ImportErrors due to missing compiled C extensions. We recommend that you try running your code from outside the source tree. If you are outside the source tree then you have a setup.py file in an unexpected directory: /opt/homebrew/lib/python3.11/site-packages
warnings.warn(

AttributeError: 'Exonize' object has no attribute '_Exonize__target_CDS_frame'. Did you mean:
'_Exonize__query_CDS_frame'?Classifying events %

seems like I missed to initiallize some attributes.

Check for exclusive pairs that are classified as “optional” (neither>0) but are actually MXE - it may just be a case where >2 duplicated exons may be sampled at that site.

This relates to the duplication mechanism, i.e., it is a tandem or a long region duplication.

If the best reciprocal hits are adjacent to each other this will mean that we're dealing with tandem duplications and not with a multi-exon duplication.