And one of the surprises from the widespread use of sequencing in the clinic is how many rare Mendelian disorders - especially neurodevelopmental ones - are caused by de novo variants

The figure below (from www.nejm.org/doi/full/10....) nicely illustrates this (3/n)

It's a very interesting question, probably the bottleneck contributes quite a bit. E.g. in Bergstrom and Pritchard's model (academic.oup.com/genetics/art... ) a tighter bottleneck leads to better mitochondrial "quality" and slows down the rate of deterioration at the population level