Show annotations of genetic elements in Sanger trace visualisation

[sebschulz1]

draft_gene_annotation_sage.pptx

The current Sanger trace visualisation (Sage) does not show annotations of genetic elements. The DNAFoundry has asked for this functionality. Showing annotations would substantially improve navigation of the visualisation results and consequently the user experience.
As discussed previously, this might require changing the source code of the visualisation software itself. It might be an interesting opportunity to reach out to the tracy developer(s) for collaboration.
Beyond the current visualisation tool (Sage), other tools (Pearl and Sabre) would benefit from this functionality too.

Annotations can be obtained form the following annotation file types:

• GenBank file (.gb, .gbk) --> PRIORITY: we will definitely need this because this can be exported from Benchling as the only annotation file type so far (conversion of such files into gtf and/or gff is probably possible but I would like to avoid this step if possible) --> more info of the GenBank format can be found here
• GTF (.gtf)
• GFF (.gff) or GFF3 (.gff3)
  --> Sebastian to look for tools that parse these files

Relevant information to be extracted from the annotation files:

• name and ID of genetic elements
• gene coordinates (string index), i.e. start and end of each genetic element on a DNA string
• strandedness, i.e. on which DNA strand the genetic element is encoded. Can be either sense (+) or antisense (-) strand. In the GenBank file this is indicated by e.g. complement when the genetic feature is on the complementary strans.

Requirements for visualisation in Sage:

• Show the genetic element below or overlapping with the DNA string using genomic coordinates (the shape is ideally a box with an arrow)
• Show the strandedness of the genetic element using the arrow direction. A genetic element on the minus strand is shown as an arrow pointing to the left whereas a genetic element on the plus strand is shown as an arrow pointing to the right)
• Show the name and ID of the genetic element

Example draft visualisation and data

• A draft concept visualisation is attached.
• Test data will be shared via a separate channel

[Skitionek]

Reading annotations probably can be done using: https://biopython.org/docs/1.75/api/Bio.SeqIO.html#Bio.SeqIO.read

[sebschulz1]

Reading annotations probably can be done using: https://biopython.org/docs/1.75/api/Bio.SeqIO.html#Bio.SeqIO.read

I also came across this. Agree, worth checking.

[Skitionek]

@sebschulz1 I am confused as in .fa .gb files has 6,576 bases while align files has 1,230. How do I convert the index from the first to the second?

[Skitionek]

I see that align result has an substring of .fa (sometimes reversed) with '-' added in random positions. Do you have knowledge on how to do the index?

[Skitionek]

Found ref_pos indicating start of subsequence from fa 🎉

[sebschulz1]

@sebschulz1 I am confused as in .fa .gb files has 6,576 bases while align files has 1,230. How do I convert the index from the first to the second?

@Skitionek The alignment of the experimental result (DNA sequence from .ab1 file) is usually shorter than the reference sequence. The alignment algorithm finds where the shorter experimentally determined DNA string aligns to the larger reference DNA string (also introducing gaps like "-" which can represent alterations in the experimentally determined DNA sequence like deletions).
The index "matching" of shorter DNA and reference DNA string should be somewhere in the results, e.g. the json file (I would also have to look where exactly).
I think your last comment above (#26 (comment)) mentions that you found it in the results? If you need additional information, please let me know.

[Skitionek]

The index "matching" of shorter DNA and reference DNA string should be somewhere in the results, e.g. the json file (I would also have to look where exactly). I think your last comment above (#26 (comment)) mentions that you found it in the results? If you need additional information, please let me know.

I did some visualisation but I still struggle to have definitive response about matching these indexes. I would use your help to find the sequence index of aligned part.
I can see for instant seemingly relevant mapping in decompose results.

I've place my work so far in here: https://github.com/biosustain/annotated-align-visualiser
Most things seems right beside missing index mapping.

Pausing work for now since I cannot get any further without meantioned mapping.

[sebschulz1]

Thanks @Skitionek! I will have another look and call a meeting later this week to discuss index matching in details. I think I don't have access to https://github.com/biosustain/annotated-align-visualiser. Can you give me access please? I can then have a more detailed look.

[Skitionek]

@sebschulz1 I've updated access so it is visible to whole IP

[Skitionek]

We've just had meeting it seems:

• confirmed that ref_pos indicates where ref seq from align file starts in original seqence
• then each position can be mapped while accounting for insertions '-'. Example:

AB-C ->
A idx: ref_pos
B idx: ref_pos + 1
C idx: ref_pos + 2

• under current assumption alignments are already reverse complemented if needed