Questions about the map_permutations and apply_permutations functions

[Jesson-mark]

Hi,
thanks for the great software to discover eQTLs!

I'm currently using tensorQTL to discovert trans-eQTLs for my RNA-seq dataset and corresponding genotypes. After looking into the documentation and source code of tensorQTL, I'm still confused by some questions.

1. Could you provide a full workflow codes for a typical trans-eQTL discovery?
   The README.md of tensorQTL only contains codes for a nominal trans-eQTL mapping but without any descriptions about map_permutations and apply_permutations functions.

2. Which algorithm does map_permutations use?
   Since tensorQTL is a reimplementation of FastQTL, however I didn't find any descriptions about trans-eQTL mapping in the paper of FastQTL. Rather I found a description of permutations in trans-eQTL mapping in the paper of QTLtools. Is that algorithm of QTLtools the same as used in map_permutations function of tensorQTL?

3. Is it needed to apply tensorqtl.calculate_qvalues for outputs produced by apply_permutations function?
   It seems that calculate_qvalues is done in the phenotype-level (as in map_cis) but the outputs of apply_permutations (or specifically map_trans) are phenotype-variant associations. If tensorqtl.calculate_qvalues is not needed here, then how to calculate the FDR of each phenotype and retain the significant trans-associations after multiple testing correction?

Best regards,
Jie Wang