A knowledge-based ingredient formulation system for chemical product development in the personal care industry

Formulation Graphs for Mapping Structure-Composition of Battery Electrolytes to Device Performance

Advanced computational methods are being actively sought to address the challenges associated with the discovery and development of new combinatorial materials, such as formulations. A widely adopted approach involves domain-informed high-throughput screening of individual components that can be combined together to form a formulation. This manages to accelerate the discovery of new compounds for a target application but still leaves the process of identifying the right "formulation" from the shortlisted chemical space largely a laboratory experiment-driven process. We report a deep learning model, the Formulation Graph Convolution Network (F-GCN), that can map the structure-composition relationship of the formulation constituents to the property of liquid formulation as a whole. Multiple GCNs are assembled in parallel that featurize formulation constituents domain-intuitively on the fly. The resulting molecular descriptors are scaled based on the respective constituent's molar percentage in the formulation, followed by integration into a combined formulation descriptor that represents the complete formulation to an external learning architecture. The use case of the proposed formulation learning model is demonstrated for battery electrolytes by training and testing it on two exemplary data sets representing electrolyte formulations vs battery performance: one data set is sourced from the literature about Li/Cu half-cells, while the other is obtained by lab experiments related to lithium-iodide full-cell chemistry. The model is shown to predict performance metrics such as Coulombic efficiency (CE) and specific capacity of new electrolyte formulations with the lowest reported errors. The best-performing F-GCN model uses molecular descriptors derived from molecular graphs (GCNs) that are informed with HOMO-LUMO and electric moment properties of the molecules using a knowledge transfer technique.

Artificial Intelligence in Pharmaceutical Technology and Drug Delivery Design

Artificial intelligence (AI) has emerged as a powerful tool that harnesses anthropomorphic knowledge and provides expedited solutions to complex challenges. Remarkable advancements in AI technology and machine learning present a transformative opportunity in the drug discovery, formulation, and testing of pharmaceutical dosage forms. By utilizing AI algorithms that analyze extensive biological data, including genomics and proteomics, researchers can identify disease-associated targets and predict their interactions with potential drug candidates. This enables a more efficient and targeted approach to drug discovery, thereby increasing the likelihood of successful drug approvals. Furthermore, AI can contribute to reducing development costs by optimizing research and development processes. Machine learning algorithms assist in experimental design and can predict the pharmacokinetics and toxicity of drug candidates. This capability enables the prioritization and optimization of lead compounds, reducing the need for extensive and costly animal testing. Personalized medicine approaches can be facilitated through AI algorithms that analyze real-world patient data, leading to more effective treatment outcomes and improved patient adherence. This comprehensive review explores the wide-ranging applications of AI in drug discovery, drug delivery dosage form designs, process optimization, testing, and pharmacokinetics/pharmacodynamics (PK/PD) studies. This review provides an overview of various AI-based approaches utilized in pharmaceutical technology, highlighting their benefits and drawbacks. Nevertheless, the continued investment in and exploration of AI in the pharmaceutical industry offer exciting prospects for enhancing drug development processes and patient care.

In Silico Prediction of Food Properties: A Multiscale Perspective

Several open software packages have popularized modeling and simulation strategies at the food product scale. Food processing and key digestion steps can be described in 3D using the principles of continuum mechanics. However, compared to other branches of engineering, the necessary transport, mechanical, chemical, and thermodynamic properties have been insufficiently tabulated and documented. Natural variability, accented by food evolution during processing and deconstruction, requires considering composition and structure-dependent properties. This review presents practical approaches where the premises for modeling and simulation start at a so-called “microscopic” scale where constituents or phase properties are known. The concept of microscopic or ground scale is shown to be very flexible from atoms to cellular structures. Zooming in on spatial details tends to increase the overall cost of simulations and the integration over food regions or time scales. The independence of scales facilitates the reuse of calculations and makes multiscale modeling capable of meeting food manufacturing needs. On one hand, new image-modeling strategies without equations or meshes are emerging. On the other hand, complex notions such as compositional effects, multiphase organization, and non-equilibrium thermodynamics are naturally incorporated in models without linearization or simplifications. Multiscale method’s applicability to hierarchically predict food properties is discussed with comprehensive examples relevant to food science, engineering and packaging. Entropy-driven properties such as transport and sorption are emphasized to illustrate how microscopic details bring new degrees of freedom to explore food-specific concepts such as safety, bioavailability, shelf-life and food formulation. Routes for performing spatial and temporal homogenization with and without chemical details are developed. Creating a community sharing computational codes, force fields, and generic food structures is the next step and should be encouraged. This paper provides a framework for the transfer of results from other fields and the development of methods specific to the food domain.

Integrated “Generate, Make, and Test” for Formulated Products using Knowledge Graphs

In the multi-billion dollar formulated product industry, state of the art continues to rely heavily on experts during the “generate, make and test” steps of formulation design. We propose automation aids to each step with a knowledge graph of relevant information as the central artifact. The generate step usually focuses on coming up with new recipes for intended formulation. We propose to aid the experts who generally carry out this step manually by providing a recommendation system and a templating system on top of the knowledge graph. Using the former, the expert can create a recipe from scratch using historical formulations and related data. With the latter, the expert starts with a recipe template created by our system and substitutes the requisite constituents to form a recipe. In the current state of practice, the three steps mentioned above operate in a fragmented manner wherein observations from one step do not aid other steps in a streamlined manner. Instead of manually operated labs for the make and test steps, we assume automated or robotic labs and in-silico testing, respectively. Using two formulations, namely face cream and an exterior coating, we show how the knowledge graph may help integrate and streamline the communication between the generate, the make, and the test steps. Our initial exploration shows considerable promise.

Predicting Transdermal Uptake of Phthalates and a Paraben from Cosmetic Cream Using the Measured Fugacity

Transdermal uptake models compliment in vitro and in vivo experiments in assessing risk of environmental exposures to semivolatile organic compounds (SVOCs). A key parameter for mechanistic models is the chemical driving force for mass transfer from environmental media to human skin. In this research, we measure this driving force in the form of fugacity for chemicals in cosmetic cream and use it to model uptake from cosmetics as a surrogate for condensed environmental media. A simple cosmetic cream, containing no target analytes, was mixed with diethyl phthalate (DEP), di-n-butyl phthalate (DnBP), and butyl paraben (BP) and diluted to make creams with concentrations ranging from 0.025% to 6%. The fugacity, relative to the pure compound, was measured using solid-phase micro extraction (SPME). We found that the relationship between the concentration and fugacity is highly nonlinear. The relative fugacity of the chemicals for a 2% w/w formulation was used in a diffusion-based model to predict transdermal uptake of each chemical and was compared with excretion data from a prior human subject study with the same formulation. Dynamic simulations of excretion are generally consistent with the results of the human subject experiment but sensitive to the input parameters, especially the time between cream application and showering.

Information Extraction and Graph Representation for the Design of Formulated Products

Formulated products like cosmetics, personal and household care, and pharmaceutical products are ubiquitous in everyday life. The multi-billion-dollar formulated products industry depends primarily on experiential knowledge for the design of new products. Vast knowledge of formulation ingredients and recipes exists in offline and online resources. Experts often use rudimentary searches over this data to find ingredients and construct recipes. This state of the art leads to considerable time to market and cost. We present an approach for formulated product design that enables extraction, storage, and non-trivial search of details required for product variant generation. Our contributions are threefold. First, we show how various information extraction techniques can be used to extract ingredients and recipe actions from textual sources. Second, we describe how to store this highly connected information as a graph database with an extensible domain model. And third, we demonstrate an aid to experts in putting together a new product based on non-trivial search. In an ongoing proof of concept, we use 410 formulations of various cosmetic creams to demonstrate these capabilities with promising results.

Safety and health index development for formulated product design: Paint formulation

Over the years, safety and health effects among consumers due to the exposure of formulated products have been reported. Thus, there is a need for systematic methodologies to assess the safety and health effects of the candidate’s ingredients in the early stages of formulated product design. Therefore, an index-based methodology was proposed to assess the safety and health effects in formulated product design. Product Safety and Health Index (PSHI) highlights the health sub-indexes based on the exposure routes including eye, inhalation, ingestion, and dermal. Each exposure route has its corresponding health sub-indexes that have to be applied. There are also new sub-indexes introduced for ingestion and dermal exposure. A case study on paint formulation was used to illustrate the developed methodology. The results show that the newly proposed index is able to identify hazardous chemical ingredient(s) with its corresponding adverse safety and health effects.

Advances in chemical product design

The nature of chemical product design problems is diverse and multidisciplinary. It involves many design issues such as project management, market study, product design, process design, and economic analysis for better organizing the product design project and achieving better products. This article provides an overview of chemical product design with a multidisciplinary hierarchical framework including all the design issues and tasks. Each of the design issues and tasks are introduced and discussed, methods and tools are summarized and compared, challenges and perspectives are presented to help the chemical product design researchers on finding more novel, innovative and sustainable products, by the combined effort from academia and industry to develop a systematic generic framework, and tools including product simulator, process simulator, database manager, modeling tool, and templates for design problems.

Knowledge management in the QbD paradigm: manufacturing of biotech therapeutics

In the quality by design (QbD) paradigm, global regulatory agencies have introduced the concepts of quality risk management and knowledge management (KM) as enablers for an enhanced pharmaceutical quality system. Although the concept of quality risk management has been well elucidated in the literature, the topic of KM has received relatively scant attention. In this paper we present an opinion on KM in the QbD paradigm as it relates to the manufacturing of biotech therapeutic products. Both academic and industrial viewpoints have been considered and key gaps have been elucidated. The authors conclude that there is an urgent need for the biotech industry to create efficient KM approaches if they wish to be successful in QbD implementation.