When doing Illumina 2x150bp sequencing of genomic DNA, and after aligning the reads to GRCh38, does the percentage of the non-N fraction of the human genome as MAPQ=0 depend on the insert sizes of the genomic fragments?
This is, for two identical samples with identical final coverage, sample A having an average insert size of 250bp and sample B having an average of 450bp, would the fraction of MAPQ=0 change between the two?
Yes, as a general rule of thumb mappability increases with insert size (up to a limit) and read length. Whether this will actually occur in a given case will depend more on how randomly the sequencing samples from the genome to begin with (i.e., if the library prep happens to select for/against high mappability regions then the insert size won't much matter). The basic reason behind this is that if one end of a read aligns to a repetitive element that the likelihood of being able to use the other end as an anchor will increase a bit if it's further away. Now having said that, Illumina sequencers in particular like a certain fragment length for efficient sequencing, so you can't go too crazy on increasing the insert sizes and still get decent output (unless you modify your library prep).
This is, for two identical samples with identical final coverage, sample A having an average insert size of 250bp and sample B having an average of 450bp, would the fraction of MAPQ=0 change between the two?
For human, I expect sample B will have fewer MAPQ=0 reads (I don't have concrete data to prove). A large part of human genome consists of ALU repeats. The average size of an ALU is ~320bp. If your insert size is ~450bp, you can map reads into most of isolated ALUs in principle. With ~250bp insert, fragments in the middle of ALUs won't be mappable uniquely.
Note that to see the difference, you need to use a mapper that sufficiently takes the advantage of paired-end information.
