Integrating Machine Learning and Color Chemistry: Developing a High-School Curriculum toward Real-World Problem-Solving

Artificial intelligence (AI) is rapidly transforming our world, making it imperative to educate the next generation about both the potential benefits and challenges associated with AI. This study presents a cross-disciplinary curriculum that connects AI and chemistry disciplines in the high school classroom. Particularly, we leverage machine learning (ML), an important and simple application of AI to instruct students to build an ML-based virtual pH meter for high-precision pH read-outs. We used a "codeless" and free ML neural network building software - Orange, along with a simple chemical topic of pH to show the connection between AI and chemistry for high-schoolers who might have rudimentary backgrounds in both disciplines. The goal of this curriculum is to promote student interest and drive in the analytical chemistry domain and offer insights into how the interconnection between chemistry and ML can benefit high-school students in science learning. The activity involves students using pH strips to measure the pH of various solutions with local relevancy and then building an ML neural network model to predict the pH value based on color changes of pH strips. The integrated curriculum increased student interest in chemistry and ML and demonstrated the relevance of science to their daily lives and global issues. This approach is transformative in developing a broad spectrum of integration topics between chemistry and ML and understanding their global impacts.

AI-driven robotic chemist for autonomous synthesis of organic molecules

The automation of organic compound synthesis is pivotal for expediting the development of such compounds. In addition, enhancing development efficiency can be achieved by incorporating autonomous functions alongside automation. To achieve this, we developed an autonomous synthesis robot that harnesses the power of artificial intelligence (AI) and robotic technology to establish optimal synthetic recipes. Given a target molecule, our AI initially plans synthetic pathways and defines reaction conditions. It then iteratively refines these plans using feedback from the experimental robot, gradually optimizing the recipe. The system performance was validated by successfully determining synthetic recipes for three organic compounds, yielding that conversion rates that outperform existing references. Notably, this autonomous system is designed around batch reactors, making it accessible and valuable to chemists in standard laboratory settings, thereby streamlining research endeavors.

Transfer Learning of Full Molecular Weight Distributions via High-Throughput Computer-Controlled Polymerization

The skew and shape of the molecular weight distribution (MWD) of polymers have a significant impact on polymer physical properties. Standard summary metrics statistically derived from the MWD only provide an incomplete picture of the polymer MWD. Machine learning (ML) methods coupled with high-throughput experimentation (HTE) could potentially allow for the prediction of the entire polymer MWD without information loss. In our work, we demonstrate a computer-controlled HTE platform that is able to run up to 8 unique variable conditions in parallel for the free radical polymerization of styrene. The segmented-flow HTE system was equipped with an inline Raman spectrometer and offline size exclusion chromatography (SEC) to obtain time-dependent conversion and MWD, respectively. Using ML forward models, we first predict monomer conversion, intrinsically learning varying polymerization kinetics that change for each experimental condition. In addition, we predict entire MWDs including the skew and shape as well as SHAP analysis to interpret the dependence on reagent concentrations and reaction time. We then used a transfer learning approach to use the data from our high-throughput flow reactor to predict batch polymerization MWDs with only three additional data points. Overall, we demonstrate that the combination of HTE and ML provides a high level of predictive accuracy in determining polymerization outcomes. Transfer learning can allow exploration outside existing parameter spaces efficiently, providing polymer chemists with the ability to target the synthesis of polymers with desired properties.

Readily Accessible High-Throughput Experimentation: A General Protocol for the Preparation of ChemBeads and EnzyBeads

Chemical- and enzyme-coated beads (ChemBeads and EnzyBeads) were introduced recently as a universal strategy for the accurate dispensing of various solids in submilligram quantities using automated instrumentation or manual dispensing. The coated beads are prepared using a resonant acoustic mixer (RAM)-an instrument that may be available only to well-established facilities. In this study, we evaluated alternative coating methods for preparing ChemBeads and EnzyBeads without the use of a RAM. We also evaluated the effects of bead sizes on loading accuracy using 4 coating methods and 12 solids (9 chemicals and 3 enzymes) as test subjects. While our original RAM coating method is the most versatile for the broadest range of solids, high-quality ChemBeads and EnzyBeads that are suitable for high-throughput experimentation can be prepared using alternative methods. These results should make ChemBeads and EnzyBeads readily accessible as the core technology for setting up high-throughput experimentation platforms.

Determining Phase Separation Dynamics with an Automated Image Processing Algorithm

The problems of extracting products efficiently from reaction workups are often overlooked. Issues such as emulsions and rag layer formation can cause long separation times and slow production, thus resulting in manufacturing inefficiencies. To better understand science within this area and to support process development, an image processing methodology has been developed that can automatically track the interface between liquid-liquid phases and provide a quantitative measure of the separation rate of two immiscible liquids. The algorithm is automated and has been successfully applied to 29 cases. Its robustness has been demonstrated with a variety of different liquid mixtures that exhibit a wide range of separation behavior-making such an algorithm suited to high-throughput experimentation. The information gathered from applying the algorithm shows how issues resulting from poor separations can be detected early in process development.

The case for data science in experimental chemistry: examples and recommendations

The physical sciences community is increasingly taking advantage of the possibilities offered by modern data science to solve problems in experimental chemistry and potentially to change the way we design, conduct and understand results from experiments. Successfully exploiting these opportunities involves considerable challenges. In this Expert Recommendation, we focus on experimental co-design and its importance to experimental chemistry. We provide examples of how data science is changing the way we conduct experiments, and we outline opportunities for further integration of data science and experimental chemistry to advance these fields. Our recommendations include establishing stronger links between chemists and data scientists; developing chemistry-specific data science methods; integrating algorithms, software and hardware to 'co-design' chemistry experiments from inception; and combining diverse and disparate data sources into a data network for chemistry research.

Automation isn't automatic

Automation has become an increasingly popular tool for synthetic chemists over the past decade. Recent advances in robotics and computer science have led to the emergence of automated systems that execute common laboratory procedures including parallel synthesis, reaction discovery, reaction optimization, time course studies, and crystallization development. While such systems offer many potential benefits, their implementation is rarely automatic due to the highly specialized nature of synthetic procedures. Each reaction category requires careful execution of a particular sequence of steps, the specifics of which change with different conditions and chemical systems. Careful assessment of these critical procedural requirements and identification of the tools suitable for effective experimental execution are key to developing effective automation workflows. Even then, it is often difficult to get all the components of an automated system integrated and operational. Data flows and specialized equipment present yet another level of challenge. Unfortunately, the pain points and process of implementing automated systems are often not shared or remain buried deep in the SI. This perspective provides an overview of the current state of automation of synthetic chemistry at the benchtop scale with a particular emphasis on core considerations and the ensuing challenges of deploying a system. Importantly, we aim to reframe automation as decidedly not automatic but rather an iterative process that involves a series of careful decisions (both human and computational) and constant adjustment.

Learning the structure-activity relationship (SAR) of the Wittig reaction from genetically-encoded substrates

The Wittig reaction can be used for late stage functionalization of proteins and peptides to ligate glycans, pharmacophores, and many other functionalities. In this manuscript, we modified 160 000 N-terminal glyoxaldehyde peptides displayed on phage with the Wittig reaction by using a biotin labeled ylide under conditions that functionalize only 1% of the library population. Deep-sequencing of the biotinylated and input populations estimated the rate of conversion for each sequence. This “deep conversion” (DC) from deep sequencing correlates with rate constants measured by HPLC. Peptide sequences with fast and slow reactivity highlighted the critical role of primary backbone amides (N–H) in accelerating the rate of the aqueous Wittig reaction. Experimental measurement of reaction rates and density functional theory (DFT) computation of the transition state geometries corroborated this relationship. We also collected deep-sequencing data to build structure–activity relationship (SAR) models that can predict the DC value of the Wittig reaction. By using these data, we trained two classifier models based on gradient boosted trees. These classifiers achieved area under the ROC (receiver operating characteristic) curve (ROC AUC) of 81.2 ± 0.4 and 73.7 ± 0.8 (90–92% accuracy) in determining whether a sequence belonged to the top 5% or the bottom 5% in terms of its reactivity. This model can suggest new peptides never observed experimentally with ‘HIGH’ or ‘LOW’ reactivity. Experimental measurement of reaction rates for 11 new sequences corroborated the predictions for 8 of them. We anticipate that phage-displayed peptides and related mRNA or DNA-displayed substrates can be employed in a similar fashion to study the substrate scope and mechanisms of many other chemical reactions.

160 000 peptides displayed on phage were subjected to the Wittig reaction with a biotinylated ylide. Deep-sequencing estimated the conversion rate for each sequence and unveiled the relationship between sequences and the rate of the Wittig reaction.

Automated high throughput pKa and distribution coefficient measurements of pharmaceutical compounds for the SAMPL8 blind prediction challenge

The goal of the Statistical Assessment of the Modeling of Proteins and Ligands (SAMPL) challenge is to improve the accuracy of current computational models to estimate free energy of binding, deprotonation, distribution and other associated physical properties that are useful for the design of new pharmaceutical products. New experimental datasets of physicochemical properties provide opportunities for prospective evaluation of computational prediction methods. Here, aqueous pK_a and a range of bi-phasic logD values for a variety of pharmaceutical compounds were determined through a streamlined automated process to be utilized in the SAMPL8 physical property challenge. The goal of this paper is to provide an in-depth review of the experimental methods utilized to create a comprehensive data set for the blind prediction challenge. The significance of this work involves the use of high throughput experimentation equipment and instrumentation to produce acid dissociation constants for twenty-three drug molecules, as well as distribution coefficients for eleven of those molecules.

Chemputation and the Standardization of Chemical Informatics

The explosion in the use of machine learning for automated chemical reaction optimization is gathering pace. However, the lack of a standard architecture that connects the concept of chemical transformations universally to software and hardware provides a barrier to using the results of these optimizations and could cause the loss of relevant data and prevent reactions from being reproducible or unexpected findings verifiable or explainable. In this Perspective, we describe how the development of the field of digital chemistry or chemputation, that is the universal code-enabled control of chemical reactions using a standard language and ontology, will remove these barriers allowing users to focus on the chemistry and plug in algorithms according to the problem space to be explored or unit function to be optimized. We describe a standard hardware (the chemical processing programming architecture-the ChemPU) to encompass all chemical synthesis, an approach which unifies all chemistry automation strategies, from solid-phase peptide synthesis, to HTE flow chemistry platforms, while at the same time establishing a publication standard so that researchers can exchange chemical code (χDL) to ensure reproducibility and interoperability. Not only can a vast range of different chemistries be plugged into the hardware, but the ever-expanding developments in software and algorithms can also be accommodated. These technologies, when combined will allow chemistry, or chemputation, to follow computation-that is the running of code across many different types of capable hardware to get the same result every time with a low error rate.

Experimental Design in Polymer Chemistry-A Guide towards True Optimization of a RAFT Polymerization Using Design of Experiments (DoE)

Despite the great potential of design of experiments (DoE) for efficiency and plannability in academic research, it remains a method predominantly used in industrial processes. From our perspective though, DoE additionally provides greater information gain than conventional experimentation approaches, even for more complex systems such as chemical reactions. Hence, this work presents a comprehensive DoE investigation on thermally initiated reversible addition–fragmentation chain transfer (RAFT) polymerization of methacrylamide (MAAm). To facilitate the adaptation of DoE for virtually every other polymerization, this work provides a step-by-step application guide emphasizing the biggest challenges along the way. Optimization of the RAFT system was achieved via response surface methodology utilizing a face-centered central composite design (FC-CCD). Highly accurate prediction models for the responses of monomer conversion, theoretical and apparent number averaged molecular weights, and dispersity are presented. The obtained equations not only facilitate thorough understanding of the observed system but also allow selection of synthetic targets for each individual response by prediction of the respective optimal factor settings. This work successfully demonstrates the great capability of DoE in academic research and aims to encourage fellow scientists to incorporate the technique into their repertoire of experimental strategies.

Generation and Characterization of a Library of Novel Biologically Active Functional Surfactants (Surfmers) Using Combined High-Throughput Methods

We report the first successful combination of three distinct high-throughput techniques to deliver the accelerated design, synthesis, and property screening of a library of novel, bio-instructive, polymeric, comb-graft surfactants. These three-dimensional, surface-active materials were successfully used to control the surface properties of particles by forming a unimolecular deep layer on the surface of the particles via microfluidic processing. This strategy deliberately utilizes the surfactant to both create the stable particles and deliver a desired cell-instructive behavior. Therefore, these specifically designed, highly functional surfactants are critical to promoting a desired cell response. This library contained surfactants constructed from 20 molecularly distinct (meth)acrylic monomers, which had been pre-identified by HT screening to exhibit specific, varied, and desirable bacterial biofilm inhibitory responses. The surfactant's self-assembly properties in water were assessed by developing a novel, fully automated, HT method to determine the critical aggregation concentration. These values were used as the input data to a computational-based evaluation of the key molecular descriptors that dictated aggregation behavior. Thus, this combination of HT techniques facilitated the rapid design, generation, and evaluation of further novel, highly functional, cell-instructive surfaces by application of designed surfactants possessing complex molecular architectures.

Design of an Electron-Withdrawing Benzonitrile Ligand for Ni-Catalyzed Cross-Coupling Involving Tertiary Nucleophiles

The design of new ligands for cross-coupling is essential for developing new catalytic reactions that access valuable products such as pharmaceuticals. In this report, we exploit the reactivity of nitrile-containing additives in Ni catalysis to design a benzonitrile-containing ligand for cross-coupling involving tertiary nucleophiles. Kinetic and Hammett studies are used to elucidate the role of the optimized ligand, which demonstrate that the benzonitrile moiety acts as an electron-acceptor to promote reductive elimination over β-hydride elimination and stabilize low-valent Ni. With these conditions, a protocol for decyanation-metalation and Ni-catalyzed arylation is conducted, enabling access to quaternary α-arylnitriles from disubstituted malononitriles.

Downstream processing for influenza vaccines and candidates: An update

Seasonal and pandemic influenza outbreaks present severe health and economic burdens. To overcome limitations on influenza vaccines' availability and effectiveness, researchers chase universal vaccines providing broad, long-lasting protection against multiple influenza subtypes, and including pandemic ones. Novel influenza vaccine designs are under development, in clinical trials, or reaching the market, namely inactivated, or live-attenuated virus, virus-like particles, or recombinant antigens, searching for improved effectiveness; all these bring downstream processing (DSP) new challenges. Having to deal with new influenza strains, including pandemics, requires shorter development time, driving the development of faster bioprocesses. To cope with better upstream processes, new regulatory demands for quality and safety, and cost reduction requirements, new unit operations and integrated processes are increasing DSP efficiency for novel vaccine formats. This review covers recent advances in DSP strategies of different influenza vaccine formats. Focus is given to the improvements on relevant state-of-the-art unit operations, from harvest and clarification to purification steps, ending with sterile filtration and formulation. The development of more efficient unit operations to cope with biophysical properties of the new candidates is discussed: emphasis is given to the design of new stationary phases, 3D printing approaches, and continuous processing tools, such as continuous chromatography. The impact of the production platforms and vaccine designs on the downstream operations for the different influenza vaccine formats approved for this season are highlighted.

Automation and data-driven design of polymer therapeutics

Polymers are uniquely suited for drug delivery and biomaterial applications due to tunable structural parameters such as length, composition, architecture, and valency. To facilitate designs, researchers may explore combinatorial libraries in a high throughput fashion to correlate structure to function. However, traditional polymerization reactions including controlled living radical polymerization (CLRP) and ring-opening polymerization (ROP) require inert reaction conditions and extensive expertise to implement. With the advent of air-tolerance and automation, several polymerization techniques are now compatible with well plates and can be carried out at the benchtop, making high throughput synthesis and high throughput screening (HTS) possible. To avoid HTS pitfalls often described as "fishing expeditions," it is crucial to employ intelligent and big data approaches to maximize experimental efficiency. This is where the disruptive technologies of machine learning (ML) and artificial intelligence (AI) will likely play a role. In fact, ML and AI are already impacting small molecule drug discovery and showing signs of emerging in drug delivery. In this review, we present state-of-the-art research in drug delivery, gene delivery, antimicrobial polymers, and bioactive polymers alongside data-driven developments in drug design and organic synthesis. From this insight, important lessons are revealed for the polymer therapeutics community including the value of a closed loop design-build-test-learn workflow. This is an exciting time as researchers will gain the ability to fully explore the polymer structural landscape and establish quantitative structure-property relationships (QSPRs) with biological significance.

Automated solubility screening platform using computer vision

Solubility screening is an essential, routine process that is often labor intensive. Robotic platforms have been developed to automate some aspects of the manual labor involved. However, many of the existing systems rely on traditional analytic techniques such as high-performance liquid chromatography, which require pre-calibration for each compound and can be resource consuming. In addition, automation is not typically end-to-end, requiring user intervention to move vials, establish analytical methods for each compound and interpret the raw data. We developed a closed-loop, flexible robotic system with integrated solid and liquid dosing capabilities that relies on computer vision and iterative feedback to successfully measure caffeine solubility in multiple solvents. After initial researcher input (<2 min), the system ran autonomously, screening five different solvent systems (20-80 min each). The resulting solubility values matched those obtained using traditional manual techniques.

•

We demonstrate a modular, closed-loop robotic platform for solubility screening

•

Automated solvent titration is informed by computer vision and turbidity monitoring

•

No human intervention or HPLC analysis is required during the experimental loop

•

Solubility values obtained by the system match those obtained via traditional methods

The expert evaluation systems in drug development

This study investigates the unique properties and criteria of use intellectual systems in drug delivery. The traditional methods of pharmaceutical drug development are limited in the field of usability. In connection with the above mentioned, the task of development and creation of the intellectual system allowing to choose active pharmaceutical ingredients, auxiliary substances, which will be based on the expert system and information base of knowledge in the field of pharmaceutical production and it should be of current scientific and practical relevance. The most commonly used type of intelligent systems today are expert systems, which are computer systems capable of partially replacing a highly qualified specialist in his or her field of competence through the knowledge previously obtained from him or her. That is why expert systems are accepted to be considered together with knowledge bases, which are models of specialists’ behavior in a certain field, and for decision-making procedures of logical conclusion are applied. The field of application of expert systems is absolutely unlimited. A vivid example is their use in diagnostic tasks of modern medicine.

High-Throughput Experimentation and Continuous Flow Evaluation of Nucleophilic Aromatic Substitution Reactions

Nucleophilic aromatic substitution (S_NAr) reactions were optimized using high-throughput experimentation techniques for execution under flow conditions. A total of 3072 unique reactions were evaluated with an analysis time of ∼3.5 s per reaction using a system that combines a liquid handling robot for reaction mixture preparation with desorption electrospray ionization (DESI) mass spectrometry (MS) for analysis. The reactions were performed in bulk microtiter arrays with and without incubation. In-house developed software was used to process the data and generate heat maps of the results. This information was then used to select the most promising conditions for continuous synthesis under microfluidic reactor conditions. Our results show that this HTE approach provides robust guidance for narrowing the range of conditions needed for optimization of S_NAr reactions.

High throughput analysis enables high throughput experimentation in pharmaceutical process research

High throughput experimentation has become widely used in the discovery and development of new medicines.

Uniting laboratory automation, DoE data, and modeling techniques to accelerate chemical process development

Rapid knowledge building of chemical processes with highly automated DoE (HAD) and statistical analyses and modeling.