<aside> 🔅 Use cmd/ctrl+shift+l to switch between light/dark display mode.

</aside>

Related Pull Request:

https://github.com/SynBioDex/SBOL-utilities/pull/160

Change Log:

Changes::

• Changes in the GenBank -> SBOL3 converter to support storing multiple FeatureLocations for each Feature are very straightforward:
  • Instead of parsing only the 1st FeatureLocation of Features from GenBank, I have created an array of FeatureLocations, to store all parsed FeatureLocations.
  • Also, for this direction of conversion, I have not ensured any kind of sorting for Features or FeatureLocations, and am not sure if it is even required.
• For the SBOL3 -> GenBank direction of conversion:
  1. For each Feature, feat_loc shall hold the final FeatureLocation object (which will be just FeatureLocation if there is only 1 location, else CompoundLocation, if there are multiple).
  2. More info on CompoundLocation can be found here.
  3. For each Feature, feat_loc_array will be the "flattened" list of start and end points of all locations of a feature, sorted internally:
     • Eg: For a feature having locations - [2,4], [3,9], [1,30]; the feat_loc_array will hold [1,30,2,4,3,9]
     • This will be the basis for performing sorting among different features.
  4. Also, feat_order is a dictionary that will have key-value pairs of Feature Object - feat_locs array. While performing the final sorting of features, this dictionary is used to access the feat_locs array for each feature.
  5. In case a Feature has multiple locations, it will be stored in a feat_locs array, which will be internally sorted for each Feature on the basis of each location's start and end points lexicographically before being stored in the final Feature object.
• The unit test functions have been refactored, but not as well as this PR has. I can either push the refactoring commits from that PR here or wait for it to be merged.

Pending Fixes:

• [ ] The CompoundLocation Object takes in an operator argument which can be either of "join"/"order"/"compliment"; currently, I am using "join" whenever there are multiple locations. We need to discuss how the converter should decide which one to use based on parsed feature locations' details.
• [ ] The multiple location converter works, for the most part, except for when parsing the feature with the label "c5" in the outer sequence2_modified.gb file. This feature has the location of the form:

complement(join(1603..1891,2322..2438,2538..2633,2801..2843,
                     2918..3073,3167..3247,3874..3972,4082..4637))

But once this is round-tripped to return the sequence2_modified.gb file inside the test/test_files directory, the location becomes:

complement(join(4082..4637,3874..3972,3167..3247,
                     2918..3073,2801..2843,2538..2633,2322..2438,1603..1891))

The order inside the join operator reverses, and I am struggling to understand how to deal with this issue. I think this is related to the previous point, and maybe the operator used should not be "join" here.

Discussion:

In the previous blog, I discussed the addition of a conversion method for the opposite direction of conversion, i.e., SBOL3 → GenBank. Here, I shall discuss the handling of features with multiple locations and how they will further be stored.