Discrepancy at position_in_codon within upstream/downstream regions

[XLIU-hub]

How to determine position_in_codon in upstream and downstream regions.

To retain continuous phase for the position_in_codon between upstream (u)-5′ UTR(-) or 3' UTR''(*)-downstream(d) boundary. We proposed a calculation method on position_in_codon, which is to extend the CDS reading frame upstream of the start codon and assign codon positions (value 1/2/3) to all bases in region u or - relative to the CDS frame. Under this convention, if the first base of region - corresponds to the second position in the codon, then the immediately upstream base is assigned codon position as 1 of the same projected codon. But this could cause discrepancy in u or d regions.

Example
If we have a coding sequence with the exons and CDS coordinate as

_exons = [(5, 8), (14, 20), (30, 35), (40, 44), (50, 52), (70, 72)]
_cds = (32, 43)

Following the above calculation method: crossmapper would convert coordinates from 2 to 5 into protein positions as below:

{'position': 1, 'position_in_codon': 2, 'region': 'u', 'offset': 0} # coordinate 2
{'position': 1, 'position_in_codon': 3, 'region': 'u', 'offset': 0} # coordinate 3
{'position': 1, 'position_in_codon': 1, 'region': 'u', 'offset': 0} # coordinate 4
{'position': 4, 'position_in_codon': 2, 'region': '-', 'offset': 0} # coordinate 5

In HGVS format, p.u1.1 (coordinate 4) looks adjacent to p.u1.2 (coordinate 2), but they are not on the coordinate.

Goal of the protein positions from crossmapper

• can convert from coordinate to positions and vice versa.
• deliver meaningful results
• no discrepancy between coordinate or position_in_codon

Possible solutions

• Add warning message if the region in positions are not in CDS.
• Extend from - region and add count in offset

{'position': 4, 'position_in_codon': 2, 'region': '-', 'offset': -1} # coordinate 4

• Extend from "" region and add count in offset

{'position': 1, 'position_in_codon': 1, 'region': '', 'offset': -28} # coordinate 4

[XLIU-hub]

Valid points/cases to have position_in_codon continuous:

• Protein positions outside the CDS region cannot convert into proper HGVS, p.u1.1 and p.u1.2 will not be seen in the HGVS format.
• A mutation can happen right at the start codon, if there is another start codon ahead and the position_in_codon can tell if this start codon candidate is in frame or not. This could be useful in clinics.

Agreement/Solution
Upstream or downstream regions are treated as introns, so the position_in_codon value is a constant at u and d. However, it is essential to keep track of the frame. See at #4