Prediction of Nanoparticle Sizes for Arbitrary Methacrylates Using Artificial Neuronal Networks

Abstract

Particle sizes represent one of the key factors influencing the usability and specific targeting of nanoparticles in medical applications such as vectors for drug or gene therapy. A multi-layered graph convolutional network combined with a fully connected neuronal network is presented for the prediction of the size of nanoparticles based only on the polymer structure, the degree of polymerization, and the formulation parameters. The model is capable of predicting particle sizes obtained by nanoprecipitation of different poly(methacrylates). This includes polymers the network has not been trained with, indicating the high potential for generalizability of the model. By utilizing this model, a significant amount of time and resources can be saved in formulation optimization without extensive primary testing of material properties.

Improvement of High-Throughput Experimentation Using Synthesis Robots by the Implementation of Tailor-Made Sensors

“…Consequently, the working principle and effectiveness could be studied in detail. Both polymers were synthesized by a RAFT polymerization using the standard conditions established in the Schubert group [ 35 ].…”

A Cyclodiaryliodonium NOX Inhibitor for the Treatment of Pancreatic Cancer via Enzyme‐Activatable Targeted Delivery by Sulfated Glycosaminoglycan Derivatives

“…Notably, kidney signals became distinct at early time points in 2 h, in addition to reticuloendothelial system (RES) organs of the liver and spleen, signifying that proportions of systemically administered probes simultaneously underwent rapid renal and slow hepatobiliary excretions during drug release (Figure 3i-k). Since intrinsically hydrophobic CDAI18 with low aqueous solubility was primarily excreted via the hepatobiliary route, [37,38] the partial renal clearance enabled by sGPP NPs should favorably help alleviate the systemic toxicity. The tumor retention half-life was calculated to be 5.75 ± 0.94 h. Ex vivo imaging displayed predominant distribution in the tumor, liver, spleen, and kidneys (Figure 3l,m), verifying the efficient, targeted tumor delivery in vivo.…”

Application of Digital Methods in Polymer Science and Engineering

“…[104,105] This works a bit analogously to image recognition, but each atom is assigned a molecular fingerprint that also contains which neighboring atoms are present. [106] By normalizing the size, a constant vector of the same size can be achieved, which is ideal for transferring the structure. By subsequently applying a graph convolutional network, an entire molecule can be coded, and it can be utilized as an input parameter for ML models.…”