Should rows with few n_snps_used be filtered out?

[Maciel Rodriguez]

Hi! Thank you so much for this amazing tool!

It is helping me a lot :) 

I wanted to know if I should use some columns of the SPrediXcan output, like n_snps_used, to filter out rows with too few SNPs used. 

Some of my results have just 1 or 2 SNPs used, so I was wondering if I should keep the rows with at least 2 SNPs used, to have more reliable results. 

What do you think? Should I not worry about this and use the results even if they have only 1 SNP used or is it correct to set a threshold for the n_snps_used? If so, what threshold would you recommend?

Thank you so much in advance for your kind help!