random seeds

[jimfeng9705]

For procedure 6 (local run to generate alternative conformations), can we use --num-seeds of 16 instead of the random seeds (0,...., 16)?

On ColabFold webpage, num-seeds is recommended to use, while the local run suggested to use random seed 'flag'. Why?

[sokrypton]

you can do either locally:

[jimfeng9705]

Shall I use template, or not, to generate alternative conformations? How about recycle numbers (lower or higher)?