Dear community members,
I often meet the phrase "these mutations happen most frequently at transcribed strand" in literature (organism is diploid).
How do we understand the strand of a mutation in NGS data? I am reading papers now, but I still can not grasp the idea...For me several replication cycles should totally vanish all the info about strandness, no? I understand how we can "infer" strandness in some cases (e.g. leading/lagging strands for replication next to origin-of-replication point) - but we still see mutations on both strands after the replication, so I doubt we can classify all the mutations to their strands...
UPD: this is the closest for me to understanding - but that means that we can classify them only around replication domain? What are valleys and peaks, what is "slope of 250 rtu"? https://genomebiology.biomedcentral.com/articles/10.1186/s13059-018-1509-y#Sec8
"Direction of replication Left- and right-replicating domains were taken from  where replication timing profiles were generated in six lymphoblastoid cell lines , valleys and peaks (defined as regions with a slope with a magnitude lower than 250 rtu per Mb) were removed, after which left- and right-replicating domains were defined as timing transition regions with a negative and positive slope, respectively . In the left-replicated regions, the reference strand is used as a template for the leading strand, while the opposite strand is used as a template for the lagging strand, and vice versa for the right-replicated regions. Each domain (called territory in the original source code and data) is 20 kbp wide and annotated with the direction of replication and with replication timing."