Sampling of Structure and Sequence Space of Small Protein Folds

Linsky, Thomas W.; Noble, Kyle; Tobin, Autumn R; Crow, Rachel; Carter, L. P.; Urbauer, Jeffrey L.; Baker, David; Strauch, Eva-Maria

doi:10.1101/2021.03.10.434454

Cited by 8 publications

(13 citation statements)

References 25 publications

(59 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This model was trained on a corpus of 113,681 proteins designed by experts using Rosetta ( Alford et al (2017) ) or dTERMen ( Zhou et al (2020) ) software ( Figure 2–Figure Supplement 1, Figure 2–Figure Supplement 2 ), and achieved high performance on held-out test data. The training corpus comprised the designs reported in Rocklin et al (2017) , in Linsky et al (2021) , as well as previously unpublished designs.…”

Section: Resultsmentioning

confidence: 99%

“…Library 2 consists of designs from Linsky et al (2021) , which span several topologies (4h: four-helix bundles; HHH: three-helix bundles; beta_grasp: beta-grasp fold; coil: helical bundle; ferredoxin: ferredoxin fold; thio: thioredoxin fold; fold2 and fold4: new folds from Linsky et al (2021) ). Figure 2 – Figure Supplement 2 B shows stability scores for these designs, along with scrambled controls generated from a subset of designs by randomly shuffling the entire designed sequence.…”

Section: Methodsmentioning

confidence: 99%

“…In addition to the protein designs described in Rocklin et al (2017) and Linsky et al (2021) , our training corpus included designs from five other sets of proteins built using Rosetta or dTERMen. Because these were run with different batches of reagents, in different labs, at different times, by different people, some experimental variance likely affected the data.…”

Section: Methodsmentioning

confidence: 99%

“…Although this padding was stripped from design sequences, it was retained in scrambled sequences when we trained the new USM in this study. ( B ) Protein designs from Linsky et al (2021) ; we assayed stability scores for these designs in this study. ( C ) Blueprint-based Rosetta designs.…”

Section: Figure 1figure Supplementmentioning

confidence: 99%

See 3 more Smart Citations

Large-scale design and refinement of stable proteins using sequence-only models

Singer¹,

Novotney²,

Strickland

et al. 2021

Preprint

View full text Add to dashboard Cite

Engineered proteins generally must possess a stable structure in order to achieve their designed function. Stable designs, however, are astronomically rare within the space of all possible amino acid sequences. As a consequence, many designs must be tested computationally and experimentally in order to find stable ones, which is expensive in terms of time and resources. Here we report a neural network model that predicts protein stability based only on sequences of amino acids, and demonstrate its performance by evaluating the stability of almost 200,000 novel proteins. These include a wide range of sequence perturbations, providing a baseline for future work in the field. We also report a second neural network model that is able to generate novel stable proteins. Finally, we show that the predictive model can be used to substantially increase the stability of both expert-designed and model-generated proteins.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Figure 1figure Supplementmentioning

confidence: 99%

See 2 more Smart Citations

Large-scale design and refinement of stable proteins using sequence-only models

Singer¹,

Novotney²,

Strickland

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The structure models of these scaffolds must be quite accurate so that the positioning is correct. Using fragment assembly 7 , piecewise fragment assembly 8 , and helical extension 9 , we designed a large set of miniproteins ranging in length from 50 to 65 amino acids containing larger hydrophobic cores than previous miniprotein scaffold libraries 1 , which makes the protein more stable and more tolerant to introduction of the designed binding surfaces. 84,690 scaffolds spanning 5 different topologies with structural metrics predictive of folding were encoded in large oligonucleotide arrays and 34,507 were found to be stable using a high-throughput proteolysis based protein stability assay 10 .…”

Section: By Reducing Algorithmic Complexity From O(n)mentioning

confidence: 99%

Robust de novo design of protein binding proteins from target structural information alone

Cao

Coventry

Goreshnik

et al. 2021

Preprint

View full text Add to dashboard Cite

The design of proteins that bind to a specific site on the surface of a target protein using no information other than the three-dimensional structure of the target remains an outstanding challenge. We describe a general solution to this problem which starts with a broad exploration of the very large space of possible binding modes and interactions, and then intensifies the search in the most promising regions. We demonstrate its very broad applicability by de novo design of binding proteins to 12 diverse protein targets with very different shapes and surface properties. Biophysical characterization shows that the binders, which are all smaller than 65 amino acids, are hyperstable and bind their targets with nanomolar to picomolar affinities. We succeeded in solving crystal structures of four of the binder-target complexes, and all four are very close to the corresponding computational design models. Experimental data on nearly half a million computational designs and hundreds of thousands of point mutants provide detailed feedback on the strengths and limitations of the method and of our current understanding of protein-protein interactions, and should guide improvement of both. Our approach now enables targeted design of binders to sites of interest on a wide variety of proteins for therapeutic and diagnostic applications.

show abstract

Design of protein-binding proteins from the target structure alone

Cao

Coventry

Goreshnik

et al. 2022

Nature

244

274

View full text Add to dashboard Cite

The design of proteins that bind to a specific site on the surface of a target protein using no information other than the three-dimensional structure of the target remains a challenge1–5. Here we describe a general solution to this problem that starts with a broad exploration of the vast space of possible binding modes to a selected region of a protein surface, and then intensifies the search in the vicinity of the most promising binding modes. We demonstrate the broad applicability of this approach through the de novo design of binding proteins to 12 diverse protein targets with different shapes and surface properties. Biophysical characterization shows that the binders, which are all smaller than 65 amino acids, are hyperstable and, following experimental optimization, bind their targets with nanomolar to picomolar affinities. We succeeded in solving crystal structures of five of the binder–target complexes, and all five closely match the corresponding computational design models. Experimental data on nearly half a million computational designs and hundreds of thousands of point mutants provide detailed feedback on the strengths and limitations of the method and of our current understanding of protein–protein interactions, and should guide improvements of both. Our approach enables the targeted design of binders to sites of interest on a wide variety of proteins for therapeutic and diagnostic applications.

show abstract

Sampling of Structure and Sequence Space of Small Protein Folds

Cited by 8 publications

References 25 publications

Large-scale design and refinement of stable proteins using sequence-only models

Large-scale design and refinement of stable proteins using sequence-only models

Robust de novo design of protein binding proteins from target structural information alone

Design of protein-binding proteins from the target structure alone

Contact Info

Product

Resources

About