Bug in modkit extract full, canonical_base does not match reference sequence

[awjga]

Hi,

The results of my data after running modkit extract full does not match my reference.

From my reference, position 138 and 140 is T and 139 is C.
However, from the results of modkit extract full some read that contains position139 is identified as T in the out.
I see the same issue across the whole transcript, some positions are wrongly identified as T.

Looking at some of the reads and comparing the query_kmer with the reference sequence, it seems to be caused by an indels that is unmapped to the reference.
This is part of my reference:
135-144 (starts from 1)
cgcctctggc

For my downstream analysis, can I just discard the positions that have the wrong reference?

Below is the output of my results:

<style> </style>

5f0af38a-7492-48aa-a3de-1ed2ee404383	181	138	gene	+	+	+	2	62	9	3149	3243	0.9863281	17802	13	.	TCGCCTCTGGC	T	T	FALSE	0	3149
f41ca671-d4c5-42d6-8ee6-1b954a8d964f	127	139	gene	+	+	+	9	16	22	3149	3140	0.9082031	17802	7	.	CGCCCTTGGCC	T	T	FALSE	0	3149
ad238b5c-a836-478d-b239-a65408a0af2c	128	139	gene	+	+	+	1	39	9	3149	3185	0.9042969	17802	11	.	CGCCTTCGGCC	T	T	FALSE	0	3149
bffc67f7-c72c-4a60-a7b4-cb0b857fb67d	129	139	gene	+	+	+	3	42	12	3149	3171	0.9902344	17802	9	.	CGCCTTCGGCC	T	T	FALSE	0	3149
efa15c98-4958-40a0-9aef-d730efe32625	63	140	gene	+	+	+	0	39	76	3149	3079	0.9941406	17802	31	.	GCCTCTGGCCC	T	T	FALSE	0	3149