Introduction

Yesterday and Today i was experimenting with RFDiffusion; the specific experiment i just ran was to see if i could use RFDiffusion to design a simple helical-type structure. It was a very simple experiment, the parameters where:

contigs = "80-120" #@param {type:"string"}

pdb = "" #@param {type:"string"}

iterations = 50 #@param ["25", "50", "100", "150", "200"] {type:"raw"}

hotspot = "" #@param {type:"string"}

num_designs = 2 #@param ["1", "2", "4", "8", "16", "32"] {type:"raw"}

visual = "image" #@param ["none", "image", "interactive"]

#@markdown ---

#@markdown **symmetry** settings

#@markdown ---

symmetry = "cyclic" #@param ["none", "auto", "cyclic", "dihedral"]

order = 11 #@param ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12"] {type:"raw"}

chains = "" #@param {type:"string"}

add_potential = True #@param {type:"boolean"}

Not sure this is post is important info to share, but at the moment I found it very cool that we as humans can do such a thing, it mesmerized me a bit; here’s a photo of what the predicted backbone structure looked like towards the end of the process:

(a movie of the results of another separate trial/experiment I did):

when i ran ProteinPMNN to generate the sequences based off this backbone structure, and used AlphaFold to validate them, here’s what i got:

design:0 n:0 mpnn:0.997 plddt:0.315 ptm:0.237 i_ptm:0.205 pae:18.848 i_pae:28.018 rmsd:42.927 LVAIAALLTMYGTIAAVMADRWRFGLLAIAQPLLTGSTVAVLVATNLDTGRRQLPAVFIAGPLAPALARAYQLMRLPPALAIELARRYRPEDPVLTGALILTILTADPEAQALAAELL 

design:0 n:1 mpnn:0.987 plddt:0.225 ptm:0.229 i_ptm:0.215 pae:20.983 i_pae:27.502 rmsd:50.935 ARVLEALLELLGKVAEKTADTWEHGGLIIAKPVTTGSTVGVYVSRDLDTGRRQRPVVVAAGPLADDIAEAFNIMSLEPEKAIELAKKYKPKDKVLTGSVMLHVLNMDPEAQELAEEML 

design:0 n:2 mpnn:1.071 plddt:0.248 ptm:0.233 i_ptm:0.216 pae:21.511 i_pae:27.761 rmsd:39.387 VVATAAMIALYGTIAATTAETWRFGLLSVAKPALLGSTVSVLSSVNLKTGERNLPAVHIAGPLADSLLRVFNIMSLPPERAIELAKKYKPEEKDLDLALLLHILNMDPEAAKLAEELL 

design:0 n:3 mpnn:1.045 plddt:0.231 ptm:0.218 i_ptm:0.207 pae:21.259 i_pae:27.600 rmsd:43.734 VVETAALVTAYGTIAATNNTVWEHGLLTASLPRLLGSSVSVLVSRNLETGRRNRPAVSISGPLTPELLRAFNIMSLPDELAIELAKKYKPKDKDLTGAVMLHVLQMIPEAQELAAELL 

design:0 n:4 mpnn:0.976 plddt:0.208 ptm:0.222 i_ptm:0.211 pae:21.310 i_pae:27.947 rmsd:58.867 ANIIEALLKTYGTIAATTAATWRFGLLVIAKPVLTGSTVGVLAARNLDTGETQLPAVVIAGPLADDLLRAFAIMSLPPEKAIELAKKYKPKDKDLTGAVILTILTIIPEAQELAEKLL 

design:0 n:5 mpnn:1.010 plddt:0.216 ptm:0.218 i_ptm:0.205 pae:21.703 i_pae:27.523 rmsd:50.210 LNIISALLTLYGTIAATNADVWRFGLLVVAEPVLLGSTVGVLVAVNLLTGERQLPVVAIAGPLAPDLLRAFNLMSLPPELAAEAARRYTPRDPVLTGALLLTVLTMDPALAALAAELL 

design:0 n:6 mpnn:1.030 plddt:0.317 ptm:0.244 i_ptm:0.206 pae:17.504 i_pae:27.922 rmsd:45.793 DVTIAAMIEAYGRIAEVTADTWEHGLLSVSLPPLLGSTVSVLVSRNLLTGRRNRPAGHFSGPLAPALLRVFAVMSLPPEEAIELARRLRPEDPVLAGAVLLHVLNMIPEAQALAAELL 

design:0 n:7 mpnn:0.988 plddt:0.227 ptm:0.228 i_ptm:0.215 pae:21.126 i_pae:27.464 rmsd:52.592 AVTTAAMLELLGRVAERTADRWRHGYLSAARPVLTGSTVSVFVSTDLDTGRRNLPAVHIAGPLLDDLLRAYLLMSLPPELAIEAARRYRPRDPVLTGAVLLLVLTADPEAQELAAELL 

design:1 n:0 mpnn:1.033 plddt:0.235 ptm:0.249 i_ptm:0.235 pae:22.114 i_pae:27.097 rmsd:44.179 PLLLLLEIIRFETLVVGYMVSYAEDPEMAELWAELLAATPLQALHPTYLALARRASELLRETLKKASGAIATIALTLLGFALTLLAVTGLLTLILFLTTVMRYVKAEPIDLGLGVNVG 

design:1 n:1 mpnn:0.968 plddt:0.256 ptm:0.281 i_ptm:0.274 pae:21.438 i_pae:26.985 rmsd:49.574 VALVLLEVLRHESLVVGATVTLARDPELAELWAALLALTPLQAEHPLRLAAARRIAEHLAETLKKASGATATIALFLLGFALALLAVTALLTLLVVLTTWMSNVKAEKIDLGLGVVVG 

design:1 n:2 mpnn:0.999 plddt:0.237 ptm:0.293 i_ptm:0.287 pae:22.289 i_pae:26.826 rmsd:51.748 PVLVVLEIIRFESTIVGYTISYAEDPESAELWKQLAALTPLQAELPLRFELAKRISKFLKETLEKASGAIAAIALVLLGFALTLLAVTALLTLIVALATFAKYVRAEAVDLGLGVNVG

design:1 n:3 mpnn:1.004 plddt:0.281 ptm:0.305 i_ptm:0.299 pae:20.218 i_pae:27.096 rmsd:34.176 PLLVLLAVIDFETTIVGYMVTNAEDPEMAELWKELAAFTPLQERLPLMFELAKRISEFLKKTLEKAKGAISTIVTFLLGFALTLLAVTGLLTLIIFLTTFASYVEAEKVDLGLTVVVG 

design:1 n:4 mpnn:0.998 plddt:0.273 ptm:0.343 i_ptm:0.325 pae:20.594 i_pae:25.829 rmsd:44.707 VALVLLEQIRFEALVVGYMVTYARDPESAELWAELLAHTPLQALHPLWLELARRLSEFLRETLEKAKGAIATIALALLGFALAKLAVAALLTLVAVLTLFMYYVEAEPIDLGFGVNVG 

design:1 n:5 mpnn:1.045 plddt:0.261 ptm:0.324 i_ptm:0.308 pae:19.985 i_pae:26.315 rmsd:45.084 VLLVLLEQIRFETTIVGYMVTYAEDPEMAALWEQIAAATPLQALHPALFALARRISEYLRKTLEKVSGATATIVLTLLGFALTLLAVTGLLVLIAVLTLFASHVKAEAIDLGLTVNVG 

design:1 n:6 mpnn:1.055 plddt:0.273 ptm:0.357 i_ptm:0.341 pae:20.887 i_pae:25.531 rmsd:40.872 PLLVLLEIIRFESTIVGYMVTYAKDPKMAELWAQLLKETPLQALHETRLALAKGISEYLKETLKKASGAIATIALFLLGFALTLLAVTALLTLIVFITLFMSNVKAEAVDLGLGVNVG 

design:1 n:7 mpnn:1.011 plddt:0.235 ptm:0.259 i_ptm:0.243 pae:21.659 i_pae:26.894 rmsd:43.431 VALVVLEVIRFETTIVGYTVTKAENPEAAELWKELAKLTPLREEHPLMFELADILSEYLKETLEKASGATATIALFLLGFALTLLAVTGLLTLLAVLTLFMKYVKAEPVDLGLGVVVG

And below is the display of the “best” result so yeah