bytedance / dplm

Official Implemetation of DPLM (ICML'24) - Diffusion Language Models Are Versatile Protein Learners
https://bytedance.github.io/dplm/
Apache License 2.0
77 stars 8 forks source link

Motif Scaffolding Question #13

Open abcdxs117 opened 1 month ago

abcdxs117 commented 1 month ago

Hi,

In our initial testing of scaffold generation with two idx ranges, we see that after running the results through Alphafold our top results superimpose well to the reference structure, but the motifs are in the wrong locations (despite superimposing quite well). Additionally, it looks like the distance between these two ranges are always fixed. Is there anyway to run with a flexible distance between two motif scaffolds or any suggestions on ways to ensure the motifs remain positioned properly when aligned to reference structures?