How AlphaFold2 shaped the structural coverage of the human transmembrane proteome

Abstract

AlphaFold2 (AF2) provides a 3D structure for every known or predicted protein, opening up new prospects for virtually every field in structural biology. However, working with transmembrane protein molecules pose a notorious challenge for scientists, resulting in a limited number of experimentally determined structures. Consequently, algorithms trained on this finite training set also face difficulties. To address this issue, we recently launched the TmAlphaFold database, where predicted AlphaFold2 structures are embedded into the membrane plane and a quality assessment (plausibility of the membrane-embedded structure) is provided for each prediction using geometrical evaluation. In this paper, we analyze how AF2 has improved the structural coverage of membrane proteins compared to earlier years when only experimental structures were available, and high-throughput structure prediction was greatly limited. We also evaluate how AF2 can be used to search for (distant) homologs in highly diverse protein families. By combining quality assessment and homology search, we can pinpoint protein families where AF2 accuracy is still limited, and experimental structure determination would be desirable.

A Reversible Spherical Geometric Conversion of Protein Backbone Structure Coordinate Matrix to Three Independent Vectors of ρ, θ and φ

“…Geometrically, CCS and SCS are like two sides of one coin, too. Focusing on the improvement of algorithms (e.g., neuralization [22]) alone is probably not enough, unless we flip the coin over and take a look at the other side, i.e., data and IASCS, which allow us to extract two spherical structural features (θ and ϕ) from any protein with experimentally determined structure [147][148][149][150][151]. With the redefinition of protein backbone structure with ρ, θ and ϕ here, future work may involve exploring extensions of the spherical geometric conversion method to incorporate additional structural information, such as side-chain interactions and solvent accessibility, including the design of side chain placement algorithms with improved performance.…”

Computational Characterization of Membrane Proteins as Anticancer Targets: Current Challenges and Opportunities

“…However, due to their dynamic nature, under 5% of human SLC protein structures are available-although the number is rapidly increasing thanks to cryo-EM structures [33]. Computational tools, and in particular AlphaFold, are increasingly expanding the availability of good-quality predicted protein structures, although this task is less accurate for protein families with fewer structures available for training, which is the case for membrane proteins [34].…”

Sensitive remote homology search by local alignment of small positional embeddings from protein language models

“…Because of their reliance on the construction of MSAs, profile-based methods can have high computational overhead for database construction, query preparation, or both. Protein structure searches also show higher sensitivity than sequence searches (Jambrich et al, 2023). Until recently, the utility of structure searches for protein annotation was limited by the lack of extensive reference databases and the inability to predict structures quickly and reliably for sequences lacking experimentally determined structures.…”