Insert size distribution for SV detection

[BnaT]

I'm new to bioinformatic analysis and statistics, and this is my first post.
I've got a basic question regarding the insert size distribution of paired-end reads.
Most of PEM-based SV callers assuming a normal distribution from insert size.
I have made a plot with my own data (76bp PR from MiSeq) and I see that this distribution is right skewed. So my question is, are these callers appropiate for skewed distributions? Are they performing any data transformation? I'm a bit confused,

Thanks

[Brian Bushnell]

Insert-size distributions are always skewed, particularly from size-selected libraries. The assumption of a normal distribution is made because it makes the math easy- it would be better to use the actual distribution. But in practice, the difference should not be important, at least for structural variations much larger than the mean insert size.