Salim Sazzed
Georgia Southern University, GA·Computer Science
A protein's 3D shape determines how it acts in the body, so knowing that shape is key to understanding disease and designing drugs. Cryo-electron microscopy (cryo-EM) images frozen proteins in a near-natural state, but many of the resulting 3D maps are only medium resolution.
- Fields
Computer Science
Biophysics
Biochemistry
- Topics
cryo-electron microscopy
cryo-EM
protein structure
- Grants
- 1 known grant, through Jul 2027
What they work on
- Characterize the structural inconsistencies and variability of protein secondary structures, such as helices and sheets, in medium-resolution cryo-electron microscopy (cryo-EM) maps.
- Design a deep-learning semantic segmentation model that labels secondary structures in cryo-EM maps, trained with curriculum learning that moves from easier to harder examples.
- Combine fixed structure-fitting quality scores with adaptive curriculum strategies so the model handles structures of varying quality and geometry, and release the resulting tools as open source.
More about the research
At that level of detail, it is hard to pick out the protein's secondary structures, the local folds such as helices and sheets that make up its overall shape.
Why it matters
Many cryo-EM protein maps are too blurry for current methods to reliably trace their structural building blocks. More accurate modeling of these structures could help researchers understand how proteins work and could aid drug design algorithms that depend on protein shape. Open-source release of the tools would let other structural biologists apply them to their own data.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Willy Wriggers
Old Dominion University, VA·Biophysics
Develop measures of how well an atomic model fits local regions of a cryo-electron microscopy (cryo-EM) density map, and rate maps against reliable reference structures.
Yousuf Khan
Stanford University, CA·Biophysics
Assemble RNAmass, a large RNA sequence database including viral and non-coding RNAs. Use it to train RNA-FLM1, a foundation language model, first on all sequences and then on evolutionarily conserved ones.
Joseph Santos-Sacchi
Yale University, CT·Biophysics
Determine the cryo-EM structure of prestin, the outer hair cell motor protein behind cochlear amplification, in the expanded state that voltage brings about, complementing the known contracted-state structure.
The school
Georgia Southern University: 3 labs on PI Pages, $46M a year on research, 70 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.