Is the use of conform-gt viable if the target data has only one sample?

Question

[tito]

In the page introducing the conform-gt utility, it is mentioned that target and reference data should have at least 20 samples each.

https://faculty.washington.edu/browning/conform-gt.html

Is that mandatory? Because in the beagle 5.2 docs, it is mentioned that one should probably be using conform-gt to align the target data to the reference data, and for the target data I am imputing one sample VCF files.

Thanks in advance:)

[4galaxy77]

Probably not 'mandatory' as such, but given the software relies on estimating LD and allele freqs, if you have less than 20 samples these estimates might be poor and give you incorrect results.