TEMPy-ReFF: Fine-Tuning Cryo-EM Models
The Topf group (LIV, UKE), at the Centre for Structural Systems Biology CSSB, has created new method for the refinement of cryo-EM density maps
The images of molecular complexes generated by electron cryo-microscopy (cryo-EM) are often a bit blurry and of low resolution. This makes the process of generating molecular models that reveal the structure and function of these complexes, quite difficult and time consuming. To aid researchers in this process, CSSB’s Topf group has developed TEMPy-ReFF, a new method for atomic structure refinement in cryo-EM density maps. A description of the method was recently published in Nature Communications.
Cryo-EM maps of molecular complexes have blurry regions where the existing data is incomplete. “It’s a bit like a really bad photograph with several shadows,” explains the paper’s co-first author Thomas Mulvaney “Our task is to figure out the structure of the object that has made the shadows.” By representing atomic positions as components of a Gaussian mixture model, TEMPy-ReFF helps balance the missing parts of data in the model with existing knowledge of what the individual components should look like based on chemistry. The end result of TEMPy-ReFF’s refinement is an improved model, which is often both a better fit with the data and a more faithful representation of the underlying chemistry compared with that of existing tools.
Another advantage of TEMPy-ReFF is that it uses the variances in atomic positions, which is essentially the wiggle room in the position of each atom, as a “B-factor”. Taking the B-factor into consideration, means that there are sometimes multiple solutions which fit the blurry part of a cryo-EM map equally well and TEMPy-ReFF represents this as an ensemble. “Flexibility is intrinsic in biomolecular systems and being able to account for this flexibility enables the modeling of large conformational changes” notes the paper’s other co-first author Joe Beton.
“TEMPy-ReFF’s ability to improve atomic structures modeled from cryo-EM maps will make it an important tool for the cryo-EM community,” notes Maya Topf, the paper’s corresponding author, “TEMPy-ReFF could also open up a window to many potential drug targets as its ensemble representation will enable a more reliable prediction of small-molecule docking locations.”
The TEMPy-ReFF software is freely available for academic use through: https://www.topf-group.com/tempy-reff
Original Publication:
Beton JG*, Mulvaney T*, Cragnolini T, Topf M (2024) Cryo-EM structure and B-factor refinement with ensemble representation. Nat Commun. 15(1):444. doi: 10.1038/s41467-023-44593-1. *Equal contribution
