Certain aspects of this methodology do not perform well for small Ne populations. In particular, it is known to be difficult to infer recent Ne (with say IBDNe) in cases of severe bottlenecks. Some comments:
Hard to estimate recent Ne, so hard to estimate s conditional on Ne
The scan is a deviation from null test. The asymptotic distribution may be a poor fit for small Ne populations. See Temple and Thompson (2024). In my thesis, I show that the variance in IBD rate for the selection scan is high in an Ashkenazi Jewish sample. Either there is not a sufficiently strong selection signal, or the methodology is poor for small Ne.
Certain aspects of this methodology do not perform well for small Ne populations. In particular, it is known to be difficult to infer recent Ne (with say IBDNe) in cases of severe bottlenecks. Some comments: