Data-Driven Design of Protein-Like Single-Chain Polymer Nanoparticles

Abstract

The functional structure of proteins is heavily influenced by their folding behavior. AlphaFold, a powerful artificial intelligence (AI) program trained on information from the Protein Data Bank (PDB), was developed to predict the 3D structure of proteins from its amino acid sequence. Inspired by this, we aim to elucidate structural features of synthetic single-chain polymer nanoparticles (SCNPs) based on compositional information (monomers, chain length, molecular weight, charge, and valency) by machine learning (ML). Specifically, we demonstrate the effectiveness of ML to improve the efficiency of SCNP design and uncover important polymer design attributes to mimic protein-like structural features. To start, we randomly screened over 1000 synthesized SCNPs through a combination of high-throughput dynamic light scattering (DLS) and small-angle X-ray scattering (SAXS) and compared these results to simulated protein data from the PDB. Then, utilizing evidential neural networks (ENets), we predicted, synthesized, and characterized 30 novel compact SCNPs. Incredibly, this data-driven approach yielded 58% of the predicted SCNPs with Porod exponent ≥ 3.5 as opposed to 5% of SCNPs from the random screen. Using Shapely additive explanation (SHAP) values, we further uncovered interesting contributions of monomer content on Porod exponent and radius of gyration. From this work, we have shown that an ML-guided approach proves effective for the challenging, unintuitive problem of nanoparticle design.

A fully automated platform for photoinitiated RAFT polymerization

“…44 In our own work, the inclusion of active learning with our automated platform informs polymer synthesis and characterization until the desired polymer structure-function behavior is found, such as polymers that self-assemble into single-chain polymer nanoparticles or polymers that can form polymer-protein hybrids. 6,7,45 This self-driving laboratory in which a design-build-test-learn paradigm is employed to systematically discover new polymers with specic activities will allow nonexperts to create highly complex polymer designs reliably and efficiently.…”

Data-Driven Design of Polymer-Based Biomaterials: High-throughput Simulation, Experimentation, and Machine Learning

“…In subsequent collaborative work with Gormley and co-workers, we also find that feature vectors including solely the degree of polymerization and the composition of constitutional units is sufficient to predict and distinguish among the thermal stabilities of polymer−protein hybrid systems. 17,18 Other works in the literature, such as from Upadhya et al 12 (on structural characteristics of methacrylate-based copolymers and related conjugates) and from Aldeghi and Coley 113 (on surrogate modeling of copolymer electron affinity, ionization potential, and phase behavior) echo the evident importance and utility of including polymer composition and size information directly in their feature vectors.…”

“…Numerous techniques are available for standard polymer characterization that may provide useful descriptors. For example, Upadhya et al utilized spectroscopy and light-scattering methods to measure the radius of gyration, hydrodynamic radius, and Porod exponent for over 1,400 unique single-chain nanoparticles copolymers. , Small angle scattering profiles can also be used to obtain quantitative structural or morphological characteristics of polymer solutions using analysis tools such as CREASE (computational reverse-engineering analysis for scattering experiments) developed by Jayaraman and co-workers. − In addition, high-throughput mechanical testing − and nanoindentation , can be used to rapidly probe the physical properties of bulk and film polymers . Techniques such as nuclear magnetic resonance spectroscopy can be used to extract copolymer composition.…”

“…Polymers exhibit wide-ranging properties that make them attractive for myriad technologies. − In biomaterials, polymers (both biopolymers as well as synthetic ones) can feature chemical functionalities that engender biomemetic and biointerfacing properties. − Inspired by the biological complexity present in DNA and proteins, sequence-defined polymers have significant theoretical interest and promise for medicine, nanotechnology, and information storage, − but even random copolymers, with the right balance of hydrophobic, polar, and charged electrostatic interactions, can mimic the range of conformational behaviors observed in proteins. − Synthetic copolymers can also be tailored to interface with specific proteins, modulating their enzymatic activity or protecting them in denaturing environments. − As the synthetic toolbox for polymers expands, ,,,,, so too does their capacity to enhance or mimic biological functions, giving rise to applications in catalysis, − drug delivery, − biosensors, tissue engineering, and more.…”

Sequence Patterning, Morphology, and Dispersity in Single-Chain Nanoparticles: Insights from Simulation and Machine Learning

“…Exploring and characterizing the structure–function landscape of polymeric materials is generally nontrivial given the multitude of behaviors enabled by a large chemical and architectural space. − To contend with this challenge, machine learning (ML) techniques have been increasingly utilized to probe and understand structure–function relationships in soft materials. − In the context of single polymer chains, − supervised ML models have been proven effective at relating polymer chain characteristics to average conformational behavior, thereby expediting targeted sequence- and composition-based design tasks. Meanwhile, unsupervised ML algorithms have usefully discriminated among morphological structures formed in many-chain soft materials assembly by noncovalent and supramolecular interactions. , Collective variables obtained from unsupervised ML can also form the basis for predicting and designing morphology .…”