Sample weights

[variani]

Hi,

I'm curious whether TeraStructure has an option to specify sample weights.

My target dataset is huge compared to the reference panel: 100K vs 100 samples. Combining the target and reference data together, the TeraStructure model effectively learn relationships mainly within my target dataset simply due to the sample size domination 100K >> 100.

I know other tools such as ELAI (section 2, last item bullet in their manual) implement the weights for samples (internally, no input parameters) to balance reference/target groups of samples.

Best,
Andrey