Will high MAPQ filtering introduce false negatives?

[zskey-zn]

In highly repetitive regions such as centromeres, multiple mappings can occur, and multiple mappings lead to a decrease in MAPQ. After filtering out records with MAPQ < 30, could it result in sparse or even absent coverage in regions that originally had read coverage, thereby causing the GCI to be lower than the actual value? Therefore, does the -mq parameter need to be adjusted on a case-by-case basis?

[yeeus]

Based on my experience, you can trust your long-reads without over-filtering. An MQ30 cutoff is actually quite conservative for this type of data, so there's no need to worry about being that strict.