De-risking Pharmaceutical Tablet Manufacture Through Process Understanding,Latent Variable Modeling,and Optimization Technologies

In pharmaceutical tablet manufacturing processes, a major source of disturbance affecting drug product quality is the (lot-to-lot) variability of the incoming raw materials. A novel modeling and process optimization strategy that compensates for raw material variability is presented. The approach involves building partial least squares models that combine raw material attributes and tablet process parameters and relate these to final tablet attributes. The resulting models are used in an optimization framework to then find optimal process parameters which can satisfy all the desired requirements for the final tablet attributes, subject to the incoming raw material lots. In order to de-risk the potential (lot-to-lot) variability of raw materials on the drug product quality, the effect of raw material lot variability on the final tablet attributes was investigated using a raw material database containing a large number of lots. In this way, the raw material variability, optimal process parameter space and tablet attributes are correlated with each other and offer the opportunity of simulating a variety of changes in silico without actually performing experiments. The connectivity obtained between the three sources of variability (materials, parameters, attributes) can be considered a design space consistent with Quality by Design principles, which is defined by the ICH-Q8 guidance (USDA 2006). The effectiveness of the methodologies is illustrated through a common industrial tablet manufacturing case study.     

组学时代下机器学习方法在临床决策支持中的应用

随着组学技术的不断发展,对于不同层次和类型的生物数据的获取方法日益成熟。在疾病诊治过程中会产生大量数据,通过机器学习等人工智能方法解析复杂、多维、多尺度的疾病大数据,构建临床决策支持工具,辅助医生寻找快速且有效的疾病诊疗方案是非常必要的。在此过程中,机器学习等人工智能方法的选择显得尤为重要。基于此,本文首先从类型和算法角度对临床决策支持领域中常用的机器学习等方法进行简要综述,分别介绍了支持向量机、逻辑回归、聚类算法、Bagging、随机森林和深度学习,对机器学习等方法在临床决策支持中的应用做了相应总结和分类,并对它们的优势和不足分别进行讨论和阐述,为临床决策支持中机器学习等人工智能方法的选择提供有效参考。     

Recent applications of machine learning in medicinal chemistry

In recent decades, artificial intelligence and machine learning have played a significant role in increasing the efficiency of processes across a wide spectrum of industries. When it comes to the pharmaceutical and biotechnology sectors, numerous tools enabled by advancement of computer science have been developed and are now routinely utilized. However, there are many aspects of the drug discovery process, which can further benefit from refinement of computational methods and tools, as well as improvement of accessibility of these new technologies. In this review, examples of recent developments in machine learning application are described, which have the potential to impact different parts of the drug discovery and development flow scheme. Notably, new deep learning-based approaches across compound design and synthesis, prediction of binding, activity and ADMET properties, as well as applications of genetic algorithms are highlighted.     

Decision learning about production control as machines break down in a flexible manufacturing system

During manufacturing, there are many situations that can affect production performance. Such situations include machine breakdowns, rush orders, order changes, and order delays. When such issues occur, one has to make decisions to try to maintain production efficiency. Human decisions tend to be too late and incomplete in such contingencies. Thus a system that can make better decisions in time to maintain production performance is needed. To achieve this objective, the intelligent decision system described in this paper integrates artificial intelligence, an optimization technique, and simulation to solve such problems. The decision-making logic of the intelligent decision system is described by event graphs. It imitates the manner of human thinking. Self-learning of the decision-making process is used to strengthen the decision quality. In this study, a method of rule induction is applied to build up the self-learning system. There are two subsystems included in this system. One is rule generation and the other is knowledge management. A case for machine breakdowns is presented and discussed. A series of tests designed to validate the self-learning system are presented. These demonstrate that a rule induction method is suitable for constructing the self-learning.     

Experimental Analysis of Tablet Properties for Discrete Element Modeling of an Active Coating Process

Coating of solid dosage forms is an important unit operation in the pharmaceutical industry. In recent years, numerical simulations of drug manufacturing processes have been gaining interest as process analytical technology tools. The discrete element method (DEM) in particular is suitable to model tablet-coating processes. For the development of accurate simulations, information on the material properties of the tablets is required. In this study, the mechanical parameters Young’s modulus, coefficient of restitution (CoR), and coefficients of friction (CoF) of gastrointestinal therapeutic systems (GITS) and of active-coated GITS were measured experimentally. The dynamic angle of repose of these tablets in a drum coater was investigated to revise the CoF. The resulting values were used as input data in DEM simulations to compare simulation and experiment. A mean value of Young’s modulus of 31.9 MPa was determined by the uniaxial compression test. The CoR was found to be 0.78. For both tablet–steel and tablet–tablet friction, active-coated GITS showed a higher CoF compared with GITS. According to the values of the dynamic angle of repose, the CoF was adjusted to obtain consistent tablet motion in the simulation and in the experiment. On the basis of this experimental characterization, mechanical parameters are integrated into DEM simulation programs to perform numerical analysis of coating processes.     

Protein–DNA/RNA interactions: Machine intelligence tools and approaches in the era of artificial intelligence and big data

With the development of artificial intelligence (AI) technologies and the availability of large amounts of biological data, computational methods for proteomics have undergone a developmental process from traditional machine learning to deep learning. This review focuses on computational approaches and tools for the prediction of protein – DNA/RNA interactions using machine intelligence techniques. We provide an overview of the development progress of computational methods and summarize the advantages and shortcomings of these methods. We further compiled applications in tasks related to the protein – DNA/RNA interactions, and pointed out possible future application trends. Moreover, biological sequence-digitizing representation strategies used in different types of computational methods are also summarized and discussed.     

Digital by design approach to develop a universal deep learning AI architecture for automatic chromatographic peak integration

Chromatographic data processing has garnered attention due to multiple Food and Drug Administration 483 citations and warning letters, highlighting the need for a robust technological solution. The healthcare industry has the potential to greatly benefit from the adoption of digital technologies, but the process of implementing these technologies can be slow and complex. This article presents a “Digital by Design” managerial approach, adapted from pharmaceutical quality by design principles, for designing and implementing an artificial intelligence (AI)-based solution for chromatography peak integration process in the healthcare industry. We report the use of a convolutional neural network model to predict analytical variability for integrating chromatography peaks and propose a potential GxP framework for using AI in the healthcare industry that includes elements on data management, model management, and human-in-the-loop processes. The component on analytical variability prediction has a great potential to enable Industry 4.0 objectives on real-time release testing, automated quality control, and continuous manufacturing.     

基于机器学习的药物-靶标相互作用预测*

近年来,随着计算机硬件、软件工具和数据丰度的不断突破,以机器学习为代表的人工智能技术在生物、基础医学和药学等领域的应用不断拓展和融合,极大地推动了这些领域的发展,尤其是药物研发领域的变革。其中,药物-靶标相互作用(drug-target interactions, DTI)的识别是药物研发领域中的重要难题和人工智能技术交叉融合的热门方向,研究人员在DTI预测方面做了大量的工作,构建了许多重要的数据库,开发或拓展了各类机器学习算法和工具软件。对基于机器学习的DTI预测的基本流程进行了介绍,并对利用机器学习预测DTI的研究进行了回顾,同时对不同的机器学习方法运用于DTI预测的优缺点进行了简单总结,以期对开发更加有效的预测算法和DTI预测的发展提供帮助。     

Distinct neurocognitive strategies for comprehensions of human and artificial intelligence

Although humans have inevitably interacted with both human and artificial intelligence in real life situations, it is unknown whether the human brain engages homologous neurocognitive strategies to cope with both forms of intelligence. To investigate this, we scanned subjects, using functional MRI, while they inferred the reasoning processes conducted by human agents or by computers. We found that the inference of reasoning processes conducted by human agents but not by computers induced increased activity in the precuneus but decreased activity in the ventral medial prefrontal cortex and enhanced functional connectivity between the two brain areas. The findings provide evidence for distinct neurocognitive strategies of taking others' perspective and inhibiting the process referenced to the self that are specific to the comprehension of human intelligence.     

Real-time online monitoring of insect cell proliferation and baculovirus infection using digital differential holographic microscopy and machine learning

Real-time, detailed online information on cell cultures is essential for understanding modern biopharmaceutical production processes. The determination of key parameters, such as cell density and viability, is usually based on the offline sampling of bioreactors. Gathering offline samples is invasive, has a low time resolution, and risks altering or contaminating the production process. In contrast, measuring process parameters online provides more safety for the process, has a high time resolution, and thus can aid in timely process control actions. We used online double differential digital holographic microscopy (D3HM) and machine learning to perform non-invasive online cell concentration and viability monitoring of insect cell cultures in bioreactors. The performance of D3HM and the machine learning model was tested for a selected variety of baculovirus constructs, products, and multiplicities of infection (MOI). The results show that with online holographic microscopy insect cell proliferation and baculovirus infection can be monitored effectively in real time with high resolution for a broad range of process parameters and baculovirus constructs. The high-resolution data generated by D3HM showed the exact moment of peak cell densities and temporary events caused by feeding. Furthermore, D3HM allowed us to obtain information on the state of the cell culture at the individual cell level. Combining this detailed, real-time information about cell cultures with methodical machine learning models can increase process understanding, aid in decision-making, and allow for timely process control actions during bioreactor production of recombinant proteins.     

Rapid Development and Optimization of Tablet Manufacturing Using Statistical Tools

The purpose of this paper was to develop a statistical methodology to optimize tablet manufacturing considering drug chemical and physical properties applying a crossed experimental design. The assessed model drug was dried ferrous sulphate and the variables were the hardness and the relative proportions of three excipients, binder, filler and disintegrant. Granule properties were modeled as a function of excipient proportions and tablet parameters were defined by the excipient proportion and hardness. The desirability function was applied to achieve optimal values for excipient proportions and hardness. In conclusion, crossed experimental design using hardness as the only process variable is an efficient strategy to quickly determine the optimal design process for tablet manufacturing. This method can be applied for any tablet manufacturing method.     

Implementation of chemometrics,design of experiments,and neural network analysis for prior process knowledge assessment,failure modes and effect analysis,scale-down model development,and process characterization for a chromatographic purification of Teriparatide

Process understanding and characterization forms the foundation, ensuring consistent and robust biologics manufacturing process. Using appropriate modeling tools and machine learning approaches, the process data can be monitored in real time to avoid manufacturing risks. In this article, we have outlined an approach toward implementation of chemometrics and machine learning tools (neural network analysis) to model and predict the behavior of a mixed-mode chromatography step for a biosimilar (Teriparatide) as a case study. The process development data and process knowledge was assimilated into a prior process knowledge assessment using chemometrics tools to derive important parameters critical to performance indicators (i.e., potential quality and process attributes) and to establish the severity ranking for the FMEA analysis. The characterization data of the chromatographic operation are presented alongwith the determination of the critical, key and non- key process parameters, set points, operating, process acceptance and characterized ranges. The scale-down model establishment was assessed using traditional approaches and novel approaches like batch evolution model and neural network analysis. The batch evolution model was further used to demonstrate batch monitoring through direct chromatographic data, thus demonstrating its application for continuos process verification. Assimilation of process knowledge through a structured data acquisition approach, built-in from process development to continuous process verification was demonstrated to result in a data analytics driven model that can be coupled with machine learning tools for real time process monitoring. We recommend application of these approaches with the FDA guidance on stage wise process development and validation to reduce manufacturing risks.     

Application of machine learning methods to pathogen safety evaluation in biological manufacturing processes

The production of recombinant therapeutic proteins from animal or human cell lines entails the risk of endogenous viral contamination from cell substrates and adventitious agents from raw materials and environment. One of the approaches to control such potential viral contamination is to ensure the manufacturing process can adequately clear the potential viral contaminants. Viral clearance for production of human monoclonal antibodies is achieved by dedicated unit operations, such as low pH inactivation, viral filtration, and chromatographic separation. The process development of each viral clearance step for a new antibody production requires significant effort and resources invested in wet laboratory experiments for process characterization studies. Machine learning methods have the potential to help streamline the development and optimization of viral clearance unit operations for new therapeutic antibodies. The current work focuses on evaluating the usefulness of machine learning methods for process understanding and predictive modeling for viral clearance via a case study on low pH viral inactivation.     

The evolution of primate general and cultural intelligence

There are consistent individual differences in human intelligence, attributable to a single 'general intelligence' factor, g. The evolutionary basis of g and its links to social learning and culture remain controversial. Conflicting hypotheses regard primate cognition as divided into specialized, independently evolving modules versus a single general process. To assess how processes underlying culture relate to one another and other cognitive capacities, we compiled ecologically relevant cognitive measures from multiple domains, namely reported incidences of behavioural innovation, social learning, tool use, extractive foraging and tactical deception, in 62 primate species. All exhibited strong positive associations in principal component and factor analyses, after statistically controlling for multiple potential confounds. This highly correlated composite of cognitive traits suggests social, technical and ecological abilities have coevolved in primates, indicative of an across-species general intelligence that includes elements of cultural intelligence. Our composite species-level measure of general intelligence, 'primate g(S)', covaried with both brain volume and captive learning performance measures. Our findings question the independence of cognitive traits and do not support 'massive modularity' in primate cognition, nor an exclusively social model of primate intelligence. High general intelligence has independently evolved at least four times, with convergent evolution in capuchins, baboons, macaques and great apes.     

Process Analytical Technology (PAT): batch-to-batch reproducibility of fermentation processes by robust process operational design and control

The Process Analytical Technology (PAT) initiative of the FDA is a reaction on the increasing discrepancy between current possibilities in process supervision and control of pharmaceutical production processes and its current application in industrial manufacturing processes. With rigid approval practices based on standard operational procedures, adaptations of production reactors towards the state of the art were more or less inhibited for long years. Now PAT paves the way for continuous process and product improvements through improved process supervision based on knowledge-based data analysis, "Quality-by-Design"-concepts, and, finally, through feedback control. Examples of up-to-date implementations of this concept are presented. They are taken from one key group of processes in recombinant pharmaceutical protein manufacturing, the cultivations of genetically modified Escherichia coli bacteria.     

A distributed knowledge-based approach to flexible automation: The contract net framework

This article applied distributed artificial intelligence to the real-time planning and control of flexible manufacturing systems (FMS) consisting of asynchronous manufacturing cells. A knowledge-based approach is used to determine the course of action, resource sharing, and processor assignments. Within each cell there is an embedded automatic planning system that executes dynamic scheduling and supervises manufacturing operations. Because of the decentralized control, real-time task assignments are carried out by a negotiation process among cell hosts. The negotiation process is modeled by augmented Petri nets —the combination of production rules and Petri nets—and is excuted by a distributed, rule-based algorithm.     

A pattern-directed approach to flexible manufacturing: A framework for intelligent scheduling,learning, and control

The planning, scheduling, and control of manufacturing systems can all be viewed as problem-solving activities. In flexible manufacturing systems (FMSs), the computer program carrying out these problem-solving activities must additionally be able to handle the shorter lead time, the flexibility of job routing, the multiprocessing environment, the dynamic changing states, and the versatility of machines. This article presents an artificial intelligence (AI) method to perform manufacturing problem solving. Since the method is driven by manufacturing scenarios represented by symbolic patterns, it is referred to as pattern-directed. The method is based on three AI techniques. The first is the pattern-directed inference technique to capture the dynamic nature of FMSs. The second is the nonlinear planning technique to construct schedules and assign resources. The third is the inductive learning method to generate the pattern-directed heuristics. This article focuses on solving the FMS scheduling problem. In addition, this article reports the computation results to evaluate the utility of various heuristic functions, to identify important design parameters, and to analyze the resulting computational performance in using the pattern-directed approach for manufacturing problem-solving tasks such as scheduling.     

Understanding the yeast host cell response to recombinant membrane protein production

Membrane proteins are drug targets for a wide range of diseases. Having access to appropriate samples for further research underpins the pharmaceutical industry's strategy for developing new drugs. This is typically achieved by synthesizing a protein of interest in host cells that can be cultured on a large scale, allowing the isolation of the pure protein in quantities much higher than those found in the protein's native source. Yeast is a popular host as it is a eukaryote with similar synthetic machinery to that of the native human source cells of many proteins of interest, while also being quick, easy and cheap to grow and process. Even in these cells, the production of human membrane proteins can be plagued by low functional yields; we wish to understand why. We have identified molecular mechanisms and culture parameters underpinning high yields and have consolidated our findings to engineer improved yeast host strains. By relieving the bottlenecks to recombinant membrane protein production in yeast, we aim to contribute to the drug discovery pipeline, while providing insight into translational processes.     

Comparisons of likelihood and machine learning methods of individual classification

Classification methods used in machine learning (e.g., artificial neural networks, decision trees, and k-nearest neighbor clustering) are rarely used with population genetic data. We compare different nonparametric machine learning techniques with parametric likelihood estimations commonly employed in population genetics for purposes of assigning individuals to their population of origin ("assignment tests"). Classifier accuracy was compared across simulated data sets representing different levels of population differentiation (low and high F(ST)), number of loci surveyed (5 and 10), and allelic diversity (average of three or eight alleles per locus). Empirical data for the lake trout (Salvelinus namaycush) exhibiting levels of population differentiation comparable to those used in simulations were examined to further evaluate and compare classification methods. Classification error rates associated with artificial neural networks and likelihood estimators were lower for simulated data sets compared to k-nearest neighbor and decision tree classifiers over the entire range of parameters considered. Artificial neural networks only marginally outperformed the likelihood method for simulated data (0-2.8% lower error rates). The relative performance of each machine learning classifier improved relative likelihood estimators for empirical data sets, suggesting an ability to "learn" and utilize properties of empirical genotypic arrays intrinsic to each population. Likelihood-based estimation methods provide a more accessible option for reliable assignment of individuals to the population of origin due to the intricacies in development and evaluation of artificial neural networks.