Computational Protein Design Pruning Sequence Space
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computational structure-based protein design methods, such as OSPREY, face challenges with large protein design problems, particularly at protein-protein interfaces, where enumerating and minimizing conformations and sequences become prohibitive due to the vast size of conformation space and presence of energetically unfavorable sequences and conformations.
Innovation Solution
The introduction of novel computational methods, including FRIES (Fast Removal of Inadequately Energised Sequences) and EWAK* (Energy Window Approximation to K*), which prune sequences and conformations based on partition function values and energy windows, reducing the sequence and conformation space while maintaining accurate calculations, allowing for faster and more efficient protein design.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional OSPREY algorithms are used to enumerate and minimize conformations and sequences, then accurate protein design calculations are achieved, but computational runtime becomes prohibitive due to the vast size of conformation space
Solution Approach 1:
The patent extracts and removes energetically unfavorable sequences and conformations from the search space using energy thresholds and partition function-based filtering. This extraction approach reduces the conformation space to only the most promising candidates, dramatically reducing runtime while preserving accurate design calculations for the retained sequences.
Solution Approach 2:
The patent performs preliminary filtering of sequences and conformations based on energy thresholds and partition function values before the main optimization process. This preliminary action identifies and removes unfavorable candidates in advance, reducing the computational burden during subsequent enumeration and minimization steps.
2Manufacturing precision
If the conformation space is fully enumerated to ensure optimal sequences, then design accuracy is maximized, but the computational cost becomes prohibitive
Solution Approach 1:
The patent applies local quality filtering by evaluating and ranking sequences based on their specific energy properties and partition function values. Instead of uniformly processing the entire conformation space, the method focuses computational resources on locally optimal sequences that meet energy thresholds, achieving high design quality for the most promising candidates without exhaustive enumeration.
Solution Approach 2:
The patent changes the parameter of energy thresholding and partition function filtering to dynamically adjust the search criteria. By modifying these parameters, the method can adaptively balance between exploration of new sequences and exploitation of known favorable conformations, reducing computational complexity while maintaining design quality.
3Productivity
If stochastic approaches are used for protein design, then computational speed is improved, but the quality and reliability of predicted sequences deteriorates
Solution Approach 1:
The patent introduces dynamic filtering criteria that adapt during the optimization process. The energy thresholds and partition function-based filtering are dynamically adjusted based on the current state of the search, allowing the method to maintain reliability by focusing on high-quality sequences while achieving productivity through efficient dynamic pruning of inferior candidates.
Data Source
AI summary
In some aspects, provided herein are computational methods for structure-based protein design. These new methods and incorporated algorithms speed-up computational structure-based protein design while maintaining accurate calculations, allowing for larger, previously infeasible protein designs. In another aspect, provided herein are mutant c-RAF proteins, and conjugates comprising the same. The conjugates may be used in methods of treating cancer in a subject, such as in methods of treating a KRas mutant cancer.