Bogdan Toader
@btoader.com
Applied mathematician developing computational imaging tools
#CryoET #TeamTomo
Postdoc in Scheres & Bharat labs @ MRC Laboratory of Molecular Biology, Cambridge
btoader.com
#CryoET #TeamTomo
Postdoc in Scheres & Bharat labs @ MRC Laboratory of Molecular Biology, Cambridge
btoader.com
good link:
openreview.net/forum?id=8tQ...
openreview.net/forum?id=8tQ...
All-atom inverse protein folding through discrete flow matching
The recent breakthrough of AlphaFold3 in modeling complex biomolecular interactions, including those between proteins and ligands, nucleotides, or metal ions, creates new opportunities for protein...
openreview.net
July 15, 2025 at 8:44 AM
good link:
openreview.net/forum?id=8tQ...
openreview.net/forum?id=8tQ...
Thanks! I did not try it on symmetric structures. However, preconditioning influences the speed of convergence to a minimum (good or not), rather than which minimum it converges to, so it's not clear to me whether it will avoid the problem you mention without running it.
July 12, 2025 at 5:13 PM
Thanks! I did not try it on symmetric structures. However, preconditioning influences the speed of convergence to a minimum (good or not), rather than which minimum it converges to, so it's not clear to me whether it will avoid the problem you mention without running it.
Reposted by Bogdan Toader
We also present a solution to the problem, showing how we can efficiently estimate 2nd order information and largely mitigate the poor conditioning, opening the door for efficient, SGD-based high resolution cryo-EM methods.
June 24, 2025 at 9:31 PM
We also present a solution to the problem, showing how we can efficiently estimate 2nd order information and largely mitigate the poor conditioning, opening the door for efficient, SGD-based high resolution cryo-EM methods.
Reposted by Bogdan Toader
We show the numerical conditioning of cryo-EM reconstruction is increasingly poor at high resolution such that SGD has no real hope of resolving high resolution structures.
June 24, 2025 at 9:31 PM
We show the numerical conditioning of cryo-EM reconstruction is increasingly poor at high resolution such that SGD has no real hope of resolving high resolution structures.