Targeted sequence design within the coarse-grained polymer genome

Abstract

The chemical design of polymers with target structural and/or functional properties represents a grand challenge in materials science. While data-driven design approaches are promising, success with polymers has been limited, largely due to limitations in data availability. Here, we demonstrate the targeted sequence design of single-chain structure in polymers by combining coarse-grained modeling, machine learning, and model optimization. Nearly 2000 unique coarse-grained polymers are simulated to construct and analyze machine learning models. We find that deep neural networks inexpensively and reliably predict structural properties with limited sequence information as input. By coupling trained ML models with sequential model-based optimization, polymer sequences are proposed to exhibit globular, swollen, or rod-like behaviors, which are verified by explicit simulations. This work highlights the promising integration of coarse-grained modeling with data-driven design and represents a necessary and crucial step toward more complex polymer design efforts.

Bias free multiobjective active learning for materials design and discovery

“…For a given hypothetical dispersant, we use molecular simulation techniques to evaluate our three ("experimental") key performance indicators. Although we carry out the synthesis and experiments in silico, the number of possible dispersants and the required computational time to evaluate the performance is too large for a brute-force screening of all 53 million dispersants of our coarse-grained polymer genome 24 . Therefore, also for this in silico example, we are limited by our resources and we aim to obtain our set of Pareto-optimal materials as efficiently as possible.…”

Advances in the Multi-Orthogonal Folding of Single Polymer Chains into Single-Chain Nanoparticles

“…Recent work has applied this strategy, which should be transferrable to SCNPs, to target specific molecular conformations of heteropolymer models (globular, swollen, rod-like, etc.) using the size distribution of a small number of simulated sequences [ 86 ]. As aforementioned, designability (folding into a unique stable state) might be approached by using bonding methods with directional interactions [ 53 ].…”

Chemically specific coarse‐graining of polymers: Methods and prospects

“…While this limitation can sometimes be overcome by utilizing curated experimental datasets and restricting the design space, 374,375 another emerging strategy, which has been used in diverse applications, is to use CG polymer simulations and surrogate ML predictions to guide the design process. 361,[376][377][378][379] For example, Webb et al efficiently identified copolymers with target size within a search space of O 10 8 ð Þ sequences 361 ; Jablonka employed a similar approach coupled to an active learning scheme to discover Pareto optimal polymer dispersants. 379 Because these works were demonstrative in nature, the underlying CG models were not specific to any particular polymer chemistry.…”