The successful application of modern machine learning for time series classification is often hampered by limitations in the quality and quantity of available training data. To overcome these limitations, domain knowledge can be leveraged in the form of parameterised mechanistic dynamical models, whereby time series observations may be represented as instances of a predefined class of dynamical systems. Provided the dynamical models are interpretable in terms of domain-specific variables and their dynamic interaction, the learning process becomes interpretable as well and enables the modeler to handle sparsely and irregularly sampled data naturally. However, the internal processes of a dynamical model are often only partially observed (PO). This can lead to ambiguity regarding which particular model realization best explains a given time series observation. This problem is well-known in the literature, and a dynamical model with this issue is referred to as structurally unidentifiable. Training a classifier that ignores knowledge about a structurally unidentifiable dynamical model can negatively influence classification performance. To address this issue, we employ structural identifiability (SI) analysis to explicitly relate parameter configurations that are associated with identical system outputs. Using the derived relations in classifier training, we demonstrate that this method significantly improves the classifier’s ability to generalize to unseen data on a number of example models from the biomedical domain. This effect is especially pronounced when the number of training instances is limited. Our results demonstrate the importance of SI, a topic that has received relatively little attention from the machine learning community.

Spherical multi-layered structures are prevalent in numerous biological systems and engineered applications, including tumor spheroids, layered tissues, and multi-shell nanoparticles for targeted drug delivery. Despite their widespread occurrence, there remains a gap in modeling particle propagation through these complex structures from a molecular communication (MC) perspective. This paper introduces a generalized analytical framework for modeling diffusion-based MC in multi-layered spherical environments. The framework is capable of supporting an arbitrary number of layers and flexible transmitter-receiver positioning. As an example, the detailed formulation is presented for the three-layer sphere, which is particularly relevant for different biological models such as tumor spheroids. Through this analytical model, two drug delivery scenarios for a three-layered tumor spheroid are investigated: (1) conventional drug delivery from a point source transmitter, and (2) targeted delivery utilizing pH-sensitive carriers that release their payload within the spheroid’s acidic necrotic core (i.e., innermost layer). The analytical results are validated using particle-based simulation (PBS) for configurations with short inter-layer distances. These findings advance the understanding of molecular transport mechanisms in multi-layered spheroids and provide a theoretical foundation for designing more efficient targeted drug delivery strategies.

Stochasticity plays a key role in many biological systems, necessitating the calibration of stochastic mathematical models to interpret associated data. For model parameters to be estimated reliably, it is typically the case that they must be structurally identifiable. Yet, while theory underlying structural identifiability analysis for deterministic differential equation models is highly developed, there are currently no tools for the general assessment of stochastic models. In this work, we present a differential algebra-based framework for the structural identifiability analysis of linear and a class of near-linear partially observed stochastic differential equation (SDE) models. Our framework is based on a deterministic recurrence relation that describes the dynamics of the statistical moments of the system of SDEs. From this relation, we iteratively form a series of necessarily satisfied equations involving only the observed moments, from which we are able to establish structurally identifiable parameter combinations. We demonstrate our framework for a suite of linear (two- and n-dimensional) and non-linear (two-dimensional) models. Most importantly, we define the notion of structural identifiability for SDE models and establish the effect of the initial condition on identifiability. We conclude with a discussion on the applicability and limitations of our approach, and potential future research directions in this understudied area.

Spherical multi-layered structures are prevalent in numerous biological systems and engineered applications, including tumor spheroids, layered tissues, and multi-shell nanoparticles for targeted drug delivery. Despite their widespread occurrence, there remains a gap in modeling particle propagation through these complex structures from a molecular communication (MC) perspective. This paper introduces a generalized analytical framework for modeling diffusion-based molecular communication in multi-layered spherical environments. The framework is capable of supporting an arbitrary number of layers and flexible transmitter-receiver positioning. As an example, the detailed formulation is presented for the three-layer sphere, which is particularly relevant for different biological models such as tumor spheroids. The analytical results are validated using particle-based simulation (PBS) in scenarios that have short inter-layer distances. The findings reveal that the characteristics of each layer significantly impact molecule propagation throughout the entire structure, making their consideration crucial for designing targeted therapies and optimizing drug delivery systems.

Upper-limb prostheses are either too expensive for many consumers or have a greatly simplified choice of actions. This research aims to enable an improvement to the quality of life for recipients of these devices by providing an understanding of how we use our hands in modern everyday life. To achieve this a taxonomy of grasps used in activities of daily living has been created. A portable motion capture system was developed to collect data. Thirteen participants’ hand movements were recorded (totalling 62 hours and 10 minutes of data). From these data, 38 and 22 grasps were observed from the left and right hands, respectively. The portable system effectively captured natural hand motions, which formed an updated taxonomy of grasps.

Quantifying finger kinematics can improve the authors' understanding of finger function and facilitate the design of efficient prosthetic devices while also identifying movement disorders and assessing the impact of rehabilitation interventions. Here, the authors present a study that quantifies grasps depicted in taxonomies during selected Activities of Daily Living (ADL). A single participant held a series of standard objects using specific grasps which were used to train Convolutional Neural Networks (CNN) for each of the four fingers individually. The experiment also recorded hand manipulation of objects during ADL. Each set of ADL finger kinematic data was tested using the trained CNN, which identified and quantified the grasps required to accomplish each task. Certain grasps appeared more often depending on the finger studied, meaning that even though there are physiological interdependencies, fingers have a certain degree of autonomy in performing dexterity tasks. The identified and most frequent grasps agreed with the previously reported findings, but also highlighted that an individual might have specific dexterity needs which may vary with profession and age. The proposed method can be used to identify and quantify key grasps for finger/hand prostheses, to provide a more efficient solution that is practical in their day-to-day tasks.

This paper presents parameter estimation for the passive moment and musculoskeletal scaling coefficients of a mathematical model describing the equations of motion of human finger segments. A novel musculoskeletal model was used as the actuator and parameter estimation was performed by least squares optimization of the differences between angular data from a motion capture system, and the solutions to the equations of motion of our model. The parameter estimates determined are unique to the individual and the dataset. Model fits show good agreement with filtered data R 2 = 0.98. The hope is that this model and approach can be used to determine the passive moment and musculoskeletal scaling coefficients for all types of grasps.

AbstractTyrosine kinase inhibitors (TKIs) are routinely prescribed for the treatment of non‐small cell lung cancer (NSCLC). As with all medications, patients can experience adverse events due to TKIs. Unfortunately, the relationship between many TKIs and the occurrence of certain adverse events remains unclear. There are limited in vivo studies which focus on TKIs and their effects on different regulation pathways. Many in vitro studies, however, that investigate the effects of TKIs observe additional changes, such as changes in gene activations or protein expressions. These studies could potentially help to gain greater understanding of the mechanisms for TKI induced adverse events. However, in order to utilize these pathways in a pharmacokinetic/pharmacodynamic (PK/PD) framework, an in vitro PK/PD model needs to be developed, in order to characterize the effects of TKIs in NSCLC cell lines. Through the use of ordinary differential equations, cell viability data and nonlinear mixed effects modeling, an in vitro TKI PK/PD model was developed with estimated PK and PD parameter values for the TKIs alectinib, crizotinib, erlotinib, and gefitinib. The relative standard errors for the population parameters are all less than 25%. The inclusion of random effects enabled the model to predict individual parameter values which provided a closer fit to the observed response. It is hoped that this model can be extended to include in vitro data of certain pathways that may potentially be linked with adverse events and provide a better understanding of TKI‐induced adverse events.

The study of finger biomechanics requires special tools for accurately recording finger joint data. A marker set to evaluate finger postures during activities of daily living is needed to understand finger biomechanics in order to improve prosthesis design and clinical interventions. The purpose of this study was to evaluate the reliability of a proposed hand marker set (the Warwick marker set) to capture finger kinematics using motion capture. The marker set consisted of the application of two and three marker clusters to the fingers of twelve participants who participated in the tests across two sessions. Calibration markers were applied using a custom palpation technique. Each participant performed a series of range of motion movements and held a set of objects. Intra and inter-session reliability was calculated as well as Standard Error of Measurement (SEM) and Minimal Detectable Difference (MDD). The findings showed varying levels of intra- and inter-session reliability, ranging from poor to excellent. The SEM and MDD values were lower for the intra-session range of motion and grasp evaluation. The reduced reliability can potentially be attributed to skin artifacts, differences in marker placement, and the inherent kinematic variability of finger motion. The proposed marker set shows potential to assess finger postures and analyse activities of daily living, primarily within the context of single session tests.

The most intuitive question for market access for medicinal products is the benefit/risk (B/R) balance. The B/R assessment can conceptually be divided into subquestions related to establishing efficacy and safety. There are additional layers to the B/R ratio for medical products, including questions related to dose selection, clinical and nonclinical pharmacology, and drug quality. Explicitly stating the actual questions and how they contribute to the overall B/R provides a structure that fosters better informed cross-domain discussions. There is currently no systematic approach in the regulatory setting to assess and establish the acceptability of alternative methods and data sources. In most cases, the medicinal product sponsors tend to prioritize traditional data types and methods, which are well accepted by regulators for inclusion in regulatory submissions. This, in addition to the absence of rigor in the use and validation of new data types and methods, and the limited training of assessors in related fields can lead to increased regulatory skepticism toward new data types and methods. A data-knowledge backbone is needed to mitigate the uncertainty in efficacy and safety characterization. This white paper discusses the value of explicitly redefining and restructuring the regulatory scientific decision making around the scientific question to be addressed. The ecosystem proposed is based on three pillars: (i) a repository connecting questions, data, and methods; (ii) the development and validation of high-quality standards for data and methods; and (iii) credibility assessment. The ecosystem is applied to four use cases for illustration. The need for training and regulatory guidance is also discussed.

The goal of this paper is twofold: firstly, to provide a novel mathematical model that describes the kinematic chain of motion of the human fingers based on Lagrangian mechanics with four degrees of freedom and secondly, to estimate the model parameters using data from able-bodied individuals. In the literature there are a variety of mathematical models that have been developed to describe the motion of the human finger. These models offer little to no information on the underlying mechanisms or corresponding equations of motion. Furthermore, these models do not provide information as to how they scale with different anthropometries. The data used here is generated using an experimental procedure that considers the free response motion of each finger segment with data captured via a motion capture system. The angular data collected are then filtered and fitted to a linear second-order differential approximation of the equations of motion. The results of the study show that the free response motion of the segments is underdamped across flexion/extension and ad/abduction.

Introduction: Across multiple studies, the most common serious adverse event groups that Small Cell Lung Cancer (SCLC) patients experience, whilst undergoing chemotherapy treatment, are: Blood and Lymphatic Disorders, Infections and Infestations together with Metabolism and Nutrition Disorders. The majority of the research that investigates the relationship between adverse events and SCLC patients, focuses on specific adverse events such as neutropenia and thrombocytopenia.Aim: This study aims to utilise machine learning in order to identify those patients who are at risk of developing common serious adverse event groups, as well as their specific adverse event classification grade.Methods: Data from five clinical trial studies were analysed and 12 analysis groups were formed based on the serious adverse event group and grade.Results: The best test runs for each of the models were able to produce an area under the curve (AUC) score of at least 0.714. The best model was the Blood and Lymphatic Disorder group, SAE grade 0 vs. grade 3 (best AUC = 1, sensitivity rate = 0.84, specificity rate = 0.96).Conclusion: The top features that contributed to this prediction were total bilirubin, alkaline phosphatase, and age. Future work should investigate the relationship between these features and common SAE groups.

Predicting the muscle excitations of the hand from kinematic data, exclusively, would enable the utilisation of motion capture data for the development of muscle controlled upper-limb prostheses. A method employing an existing musculoskeletal model and a selection of optimisation techniques is proposed for the prediction of muscle excitations of the hand from kinematic data. From 13 participants 62 hours and ten minutes of hand motions in activities of daily living (ADL) have been recorded, from which the functional hand shapes occurring within this time have been determined. A hybrid method utilising a gradient descent (GD) to determine the optimal initial conditions of a, then applied, particle swarm optimisation (PSO) technique is proposed as a means of predicting the muscle excitations of the hand from kinematic data. The proposed method has been applied to the found hand shapes to determine the muscle excitation of ADL. The resultant joint angles were within 16.1 degrees of that from the inputted hand shapes, with a mean correlation between outputted and desired of 0.77. The method has been shown to be applicable with real world recorded data and future modifications to the techniques utilised aims to further improve the accuracy of the output.

Mathematical modelling has made significant contributions to the optimization of the use of antimicrobial treatments. This article discusses the key processes that such mathematical modelling should attempt to capture. In particular, this article highlights that the response of the host immune system requires quantification, and this is illustrated with a novel model structure.

Missing data is a universal problem in analysing Real-World Evidence (RWE) datasets. In RWE datasets, there is a need to understand which features best correlate with clinical outcomes. In this context, the missing status of several biomarkers may appear as gaps in the dataset that hide meaningful values for analysis. Imputation methods are general strategies that replace missing values with plausible values. Using the Flatiron NSCLC dataset, including more than 35,000 subjects, we compare the imputation performance of six such methods on missing data: predictive mean matching, expectation-maximisation, factorial analysis, random forest, generative adversarial networks and multivariate imputations with tabular networks. We also conduct extensive synthetic data experiments with structural causal models. Statistical learning from incomplete datasets should select an appropriate imputation algorithm accounting for the nature of missingness, the impact of missing data, and the distribution shift induced by the imputation algorithm. For our synthetic data experiments, tabular networks had the best overall performance. Methods using neural networks are promising for complex datasets with non-linearities. However, conventional methods such as predictive mean matching work well for the Flatiron NSCLC biomarker dataset.

Circadian rhythms modulate physiological and behavioral processes of approximately 24-h periodicity. Alterations in the circadian timing system may lead to cardiovascular, metabolic or neurological diseases, cancers and sleep disorders, as well as to disruption of quality of life. Circadian rhythms can be tracked via laboratory tests measuring hormones in salivary, urinary or blood samples, which are collected in controlled environments. These tests are unsuitable for continuous monitoring in real-life, being expensive and time consuming, producing discrete information (i.e., few values per day) and requiring controlled environmental conditions (e.g., exposure to light can alter the samples). Thus, there is a need to develop non-invasive methods and tools to track circadian rhythms in real-life conditions. In this study, 10 healthy participants wore commercial medical-rated (i.e., CE-marked) wearable sensors, which continuously measured ECG, skin body temperature and physical activity for two consecutive days. Up to 10 salivary samples per day were taken and sent to a laboratory for measuring melatonin, which was used as proxy for circadian rhythm tracking. The results presented in this paper demonstrated that Heart Rate Variability (HRV) measures, physical activity and skin temperature changed significantly after the onset of melatonin. The deep-learning model presented in this study detected the onset of melatonin with 71 % accuracy, 67 % sensitivity, 75 % specificity and 77 % area under the curve (AUC). The current study concluded that deep learning could be used to track melatonin-onset in real-life, using physiological and behavioral measures monitored via wearable and easy-to-use sensors.

A structurally identifiable micro‐rate constant mechanistic model was used to describe the interaction between pitavastatin and eltrombopag, with improved goodness‐of‐fit values through comeasurement of pitavastatin and eltrombopag. Transporter association and dissociation rate constants and passive rates out of the cell were similar between pitavastatin and eltrombopag. Translocation into the cell through transporter‐mediated uptake was six times greater for pitavastatin, leading to pronounced inhibition of pitavastatin uptake by eltrombopag. The passive rate into the cell was 91 times smaller for pitavastatin compared with eltrombopag. A semimechanistic physiologically‐based pharmacokinetic (PBPK) model was developed to evaluate the potential for clinical drug–drug interactions (DDIs). The PBPK model predicted a twofold increase in the pitavastatin peak blood concentration and area under the concentration‐time curve in the presence of eltrombopag in simulated healthy volunteers. The use of structural identifiability supporting experimental design combined with robust micro‐rate constant parameter estimates and a semimechanistic PBPK model gave more informed predictions of transporter‐mediated DDIs.

Penicillin binding proteins (PBPs) catalyzing transpeptidation reactions that stabilize the peptidoglycan component of the bacterial cell wall are the targets of β-lactams, the most clinically successful antibiotics to date. However, PBP-transpeptidation enzymology has evaded detailed analysis, because of the historical unavailability of kinetically competent assays with physiologically relevant substrates and the previously unappreciated contribution of protein cofactors to PBP activity. By re-engineering peptidoglycan synthesis, we have constructed a continuous spectrophotometric assay for transpeptidation of native or near native peptidoglycan precursors and fragments by Escherichia coli PBP1B, allowing us to (a) identify recognition elements of transpeptidase substrates, (b) reveal a novel mechanism of stereochemical editing within peptidoglycan transpeptidation, (c) assess the impact of peptidoglycan substrates on β-lactam targeting of transpeptidation, and (d) demonstrate that both substrates have to be bound before transpeptidation occurs. The results allow characterization of high molecular weight PBPs as enzymes and not merely the targets of β-lactam acylation.

A dynamic model is structurally identifiable if it is possible to infer its unknown parameters by observing its output. Structural identifiability depends on the system dynamics, output, and input, as well as on the specific values of initial conditions and parameters. Here we present a symbolic method that characterizes the input that a model requires to be structurally identifiable. It determines which derivatives must be non-zero in order to have a sufficiently exciting input. Our approach considers structural identifiability as a generalization of nonlinear observability and incorporates extended Lie derivatives. The methodology assesses structural identifiability for time-varying inputs and, additionally, it can be used to determine the input profile that is required to make the parameters structurally locally identifiable. Furthermore, it is sometimes possible to replace an experiment with time-varying input with multiple experiments with constant inputs. We implement the resulting method as a MATLAB toolbox named STRIKE-GOLDD2. This tool can assist in the design of new experiments for the purpose of parameter estimation.