circRNA hybrid node not identified if the SNV is in the end of the exon

[zhuchcn]

In this fuzz test case, there is an SNV in the end of an exon of the circRNA.

exon: 1389-1402
SNV: 1400-A/G

The peptide node SLMPEVPYAGSPLCQK spans over 4 loops, but only 2 loops have the SNV.

Sequence feature locations:

# ((query start, query end), (ref start, ref_end), subgraph ID)
0: ((0, 4), (1389, 1400), 'CIRC-FAKET00000594-E4')
1: ((4, 5), (1401, 1402), 'CIRC-FAKET00000594-E4')
2: ((4, 9), (1386, 1400), '7c858d89-f439-4cd5-b...74c4c8b2e5')
3: ((9, 10), (1401, 1402), '7c858d89-f439-4cd5-b...74c4c8b2e5')
4: ((9, 15), (1386, 1402), '3193432e-4ce8-4ebd-9...8049f2d946')
5: ((15, 16), (1386, 1389), '626d0c89-dd94-4515-a...0a0bd13084')

The query start of #3 and #4 are both 9, but they have different ref start. This is because the the codon spans over the backsplice junction site.

Variants:

0: (3, 4, 'CIRC-FAKET00000594-E4')
1: (9, 10, '7c858d89-f439-4cd5-b...74c4c8b2e5')

Loop #3 does not have the variant but this is failed to be identified.

[lydiayliu]

So this is a circRNA from an extremely short exon, with SNV in the last codon of the exon

Sorry I don't think I completely understood. Basically all ORFs have the SNV but they weren't all called?

[zhuchcn]

This is basically a hybride node. For the 9th amino acid in the node, the first nucleotide is a SNV, second is the last nucleotide of loop #2, and the 3rd is the first nucleotide of loop #3. The entire loop #3 is on this node but doesn't carry this SNV.