Integrase-mediated differentiation circuits improve evolutionary stability of burdensome and toxic functions in E. coli

Advances in synthetic biology, bioengineering, and computation allow us to rapidly and reliably program cells with increasingly complex and useful functions. However, because the functions we engineer cells to perform are typically burdensome to cell growth, they can be rapidly lost due to the processes of mutation and natural selection. Here, we show that a strategy of terminal differentiation improves the evolutionary stability of burdensome functions in a general manner by realizing a reproductive and metabolic division of labor. To implement this strategy, we develop a genetic differentiation circuit in Escherichia coli using unidirectional integrase-recombination. With terminal differentiation, differentiated cells uniquely express burdensome functions driven by the orthogonal T7 RNA polymerase, but their capacity to proliferate is limited to prevent the propagation of advantageous loss-of-function mutations that inevitably occur. We demonstrate computationally and experimentally that terminal differentiation increases duration and yield of high-burden expression and that its evolutionary stability can be improved with strategic redundancy. Further, we show this strategy can even be applied to toxic functions. Overall, this study provides an effective, generalizable approach for protecting burdensome engineered functions from evolutionary degradation.

Nonlinear model of infection wavy oscillation of COVID-19 in Japan based on diffusion kinetics

The infectious propagation of SARS-CoV-2 is continuing worldwide, and specifically, Japan is facing severe circumstances. Medical resource maintenance and action limitations remain the central measures. An analysis of long-term follow-up reports in Japan shows that the infection number follows a unique wavy oscillation, increasing and decreasing over time. However, only a few studies explain the infection wavy oscillation. This study introduces a novel nonlinear mathematical model of the new infection wavy oscillation by applying the macromolecule diffusion theory. In this model, the diffusion coefficient that depends on population density gives nonlinearity in infection propagation. As a result, our model accurately simulated infection wavy oscillations, and the infection wavy oscillation frequency and amplitude were closely linked with the recovery rate of infected individuals. In conclusion, our model provides a novel nonlinear contact infection analysis framework.

Network analysis reveals that the tumor suppressor lncRNA GAS5 acts as a double-edged sword in response to DNA damage in gastric cancer

The lncRNA GAS5 acts as a tumor suppressor and is downregulated in gastric cancer (GC). In contrast, E2F1, an important transcription factor and tumor promoter, directly inhibits miR-34c expression in GC cell lines. Furthermore, in the corresponding GC cell lines, lncRNA GAS5 directly targets E2F1. However, lncRNA GAS5 and miR-34c remain to be studied in conjunction with GC. Here, we present a dynamic Boolean network to classify gene regulation between these two non-coding RNAs (ncRNAs) in GC. This is the first study to show that lncRNA GAS5 can positively regulate miR-34c in GC through a previously unknown molecular pathway coupling lncRNA/miRNA. We compared our network to several in-vivo/in-vitro experiments and obtained an excellent agreement. We revealed that lncRNA GAS5 regulates miR-34c by targeting E2F1. Additionally, we found that lncRNA GAS5, independently of p53, inhibits GC proliferation through the ATM/p38 MAPK signaling pathway. Accordingly, our results support that E2F1 is an engaging target of drug development in tumor growth and aggressive proliferation of GC, and favorable results can be achieved through tumor suppressor lncRNA GAS5/miR-34c axis in GC. Thus, our findings unlock a new avenue for GC treatment in response to DNA damage by these ncRNAs.

Drug transporters OAT1 and OAT3 have specific effects on multiple organs and gut microbiome as revealed by contextualized metabolic network reconstructions

In vitro and in vivo studies have established the organic anion transporters OAT1 (SLC22A6, NKT) and OAT3 (SLC22A8) among the main multi-specific "drug" transporters. They also transport numerous endogenous metabolites, raising the possibility of drug-metabolite interactions (DMI). To help understand the role of these drug transporters on metabolism across scales ranging from organ systems to organelles, a formal multi-scale analysis was performed. Metabolic network reconstructions of the omics-alterations resulting from Oat1 and Oat3 gene knockouts revealed links between the microbiome and human metabolism including reactions involving small organic molecules such as dihydroxyacetone, alanine, xanthine, and p-cresol-key metabolites in independent pathways. Interestingly, pairwise organ-organ interactions were also disrupted in the two Oat knockouts, with altered liver, intestine, microbiome, and skin-related metabolism. Compared to older models focused on the "one transporter-one organ" concept, these more sophisticated reconstructions, combined with integration of a multi-microbial model and more comprehensive metabolomics data for the two transporters, provide a considerably more complex picture of how renal "drug" transporters regulate metabolism across the organelle (e.g. endoplasmic reticulum, Golgi, peroxisome), cellular, organ, inter-organ, and inter-organismal scales. The results suggest that drugs interacting with OAT1 and OAT3 can have far reaching consequences on metabolism in organs (e.g. skin) beyond the kidney. Consistent with the Remote Sensing and Signaling Theory (RSST), the analysis demonstrates how transporter-dependent metabolic signals mediate organ crosstalk (e.g., gut-liver-kidney) and inter-organismal communication (e.g., gut microbiome-host).

Integrated network-based multiple computational analyses for identification of co-expressed candidate genes associated with neurological manifestations of COVID-19

'Tripartite network' (TN) and 'combined gene network' (CGN) were constructed and their hub-bottleneck and driver nodes (44 genes) were evaluated as 'target genes' (TG) to identify 21 'candidate genes' (CG) and their relationship with neurological manifestations of COVID-19. TN was developed using neurological symptoms of COVID-19 found in literature. Under query genes (TG of TN), co-expressed genes were identified using pair-wise mutual information to genes available in RNA-Seq autopsy data of frontal cortex of COVID-19 victims. CGN was constructed with genes selected from TN and co-expressed in COVID-19. TG and their connecting genes of respective networks underwent functional analyses through findings of their enrichment terms and pair-wise 'semantic similarity scores' (SSS). A new integrated 'weighted harmonic mean score' was formulated assimilating values of SSS and STRING-based 'combined score' of the selected TG-pairs, which provided CG-pairs with properties of CGs as co-expressed and 'indispensable nodes' in CGN. Finally, six pairs sharing seven 'prevalent CGs' (ADAM10, ADAM17, AKT1, CTNNB1, ESR1, PIK3CA, FGFR1) showed linkages with the phenotypes (a) directly under neurodegeneration, neurodevelopmental diseases, tumour/cancer and cellular signalling, and (b) indirectly through other CGs under behavioural/cognitive and motor dysfunctions. The pathophysiology of 'prevalent CGs' has been discussed to interpret neurological phenotypes of COVID-19.

Cross clinical-experimental-computational qualification of in silico drug trials on human cardiac purkinje cells for proarrhythmia risk prediction

The preclinical identification of drug-induced cardiotoxicity and its translation into human risk are still major challenges in pharmaceutical drug discovery. The ICH S7B Guideline and Q&A on Clinical and Nonclinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential promotes human in silico drug trials as a novel tool for proarrhythmia risk assessment. To facilitate the use of in silico data in regulatory submissions, explanatory control compounds should be tested and documented to demonstrate consistency between predictions and the historic validation data. This study aims to quantify drug-induced electrophysiological effects on in silico cardiac human Purkinje cells, to compare them with existing in vitro rabbit data, and to assess their accuracy for clinical pro-arrhythmic risk predictions. The effects of 14 reference compounds were quantified in simulations with a population of in silico human cardiac Purkinje models. For each drug dose, five electrophysiological biomarkers were quantified at three pacing frequencies, and results compared with available in vitro experiments and clinical proarrhythmia reports. Three key results were obtained: 1) In silico, repolarization abnormalities in human Purkinje simulations predicted drug-induced arrhythmia for all risky compounds, showing higher predicted accuracy than rabbit experiments; 2) Drug-induced electrophysiological changes observed in human-based simulations showed a high degree of consistency with in vitro rabbit recordings at all pacing frequencies, and depolarization velocity and action potential duration were the most consistent biomarkers; 3) discrepancies observed for dofetilide, sotalol and terfenadine are mainly caused by species differences between humans and rabbit. Taken together, this study demonstrates higher accuracy of in silico methods compared to in vitro animal models for pro-arrhythmic risk prediction, as well as a high degree of consistency with in vitro experiments commonly used in safety pharmacology, supporting the potential for industrial and regulatory adoption of in silico trials for proarrhythmia prediction.

HOXA9 has the hallmarks of a biological switch with implications in blood cancers

Blood malignancies arise from the dysregulation of haematopoiesis. The type of blood cell and the specific order of oncogenic events initiating abnormal growth ultimately determine the cancer subtype and subsequent clinical outcome. HOXA9 plays an important role in acute myeloid leukaemia (AML) prognosis by promoting blood cell expansion and altering differentiation; however, the function of HOXA9 in other blood malignancies is still unclear. Here, we highlight the biological switch and prognosis marker properties of HOXA9 in AML and chronic myeloproliferative neoplasms (MPN). First, we establish the ability of HOXA9 to stratify AML patients with distinct cellular and clinical outcomes. Then, through the use of a computational network model of MPN, we show that the self-activation of HOXA9 and its relationship to JAK2 and TET2 can explain the branching progression of JAK2/TET2 mutant MPN patients towards divergent clinical characteristics. Finally, we predict a connection between the RUNX1 and MYB genes and a suppressive role for the NOTCH pathway in MPN diseases.

Molecular pathways enhance drug response prediction using transfer learning from cell lines to tumors and patient-derived xenografts

Computational models have been successful in predicting drug sensitivity in cancer cell line data, creating an opportunity to guide precision medicine. However, translating these models to tumors remains challenging. We propose a new transfer learning workflow that transfers drug sensitivity predicting models from large-scale cancer cell lines to both tumors and patient derived xenografts based on molecular pathways derived from genomic features. We further compute feature importance to identify pathways most important to drug response prediction. We obtained good performance on tumors (AUROC = 0.77) and patient derived xenografts from triple negative breast cancers (RMSE = 0.11). Using feature importance, we highlight the association between ER-Golgi trafficking pathway in everolimus sensitivity within breast cancer patients and the role of class II histone deacetylases and interlukine-12 in response to drugs for triple-negative breast cancer. Pathway information support transfer of drug response prediction models from cell lines to tumors and can provide biological interpretation underlying the predictions, serving as a steppingstone towards usage in clinical setting.

An expanded whole-cell model of E. coli links cellular physiology with mechanisms of growth rate control

Growth and environmental responses are essential for living organisms to survive and adapt to constantly changing environments. In order to simulate new conditions and capture dynamic responses to environmental shifts in a developing whole-cell model of E. coli, we incorporated additional regulation, including dynamics of the global regulator guanosine tetraphosphate (ppGpp), along with dynamics of amino acid biosynthesis and translation. With the model, we show that under perturbed ppGpp conditions, small molecule feedback inhibition pathways, in addition to regulation of expression, play a role in ppGpp regulation of growth. We also found that simulations with dysregulated amino acid synthesis pathways provide average amino acid concentration predictions that are comparable to experimental results but on the single-cell level, concentrations unexpectedly show regular fluctuations. Additionally, during both an upshift and downshift in nutrient availability, the simulated cell responds similarly with a transient increase in the mRNA:rRNA ratio. This additional simulation functionality should support a variety of new applications and expansions of the E. coli Whole-Cell Modeling Project.

Dynamic cybergenetic control of bacterial co-culture composition via optogenetic feedback

Communities of microbes play important roles in natural environments and hold great potential for deploying division-of-labor strategies in synthetic biology and bioproduction. However, the difficulty of controlling the composition of microbial consortia over time hinders their optimal use in many applications. Here, we present a fully automated, high-throughput platform that combines real-time measurements and computer-controlled optogenetic modulation of bacterial growth to implement precise and robust compositional control of a two-strain E. coli community. In addition, we develop a general framework for dynamic modeling of synthetic genetic circuits in the physiological context of E. coli and use a host-aware model to determine the optimal control parameters of our closed-loop compositional control system. Our platform succeeds in stabilizing the strain ratio of multiple parallel co-cultures at arbitrary levels and in changing these targets over time, opening the door for the implementation of dynamic compositional programs in synthetic bacterial communities.

Reconstruction of a catalogue of genome-scale metabolic models with enzymatic constraints using GECKO 2.0

Genome-scale metabolic models (GEMs) have been widely used for quantitative exploration of the relation between genotype and phenotype. Streamlined integration of enzyme constraints and proteomics data into such models was first enabled by the GECKO toolbox, allowing the study of phenotypes constrained by protein limitations. Here, we upgrade the toolbox in order to enhance models with enzyme and proteomics constraints for any organism with a compatible GEM reconstruction. With this, enzyme-constrained models for the budding yeasts Saccharomyces cerevisiae, Yarrowia lipolytica and Kluyveromyces marxianus are generated to study their long-term adaptation to several stress factors by incorporation of proteomics data. Predictions reveal that upregulation and high saturation of enzymes in amino acid metabolism are common across organisms and conditions, suggesting the relevance of metabolic robustness in contrast to optimal protein utilization as a cellular objective for microbial growth under stress and nutrient-limited conditions. The functionality of GECKO is expanded with an automated framework for continuous and version-controlled update of enzyme-constrained GEMs, also producing such models for Escherichia coli and Homo sapiens. In this work, we facilitate the utilization of enzyme-constrained GEMs in basic science, metabolic engineering and synthetic biology purposes.

Artificial neural networks enable genome-scale simulations of intracellular signaling

Mammalian cells adapt their functional state in response to external signals in form of ligands that bind receptors on the cell-surface. Mechanistically, this involves signal-processing through a complex network of molecular interactions that govern transcription factor activity patterns. Computer simulations of the information flow through this network could help predict cellular responses in health and disease. Here we develop a recurrent neural network framework constrained by prior knowledge of the signaling network with ligand-concentrations as input and transcription factor-activity as output. Applied to synthetic data, it predicts unseen test-data (Pearson correlation r = 0.98) and the effects of gene knockouts (r = 0.8). We stimulate macrophages with 59 different ligands, with and without the addition of lipopolysaccharide, and collect transcriptomics data. The framework predicts this data under cross-validation (r = 0.8) and knockout simulations suggest a role for RIPK1 in modulating the lipopolysaccharide response. This work demonstrates the feasibility of genome-scale simulations of intracellular signaling.

Dynamic modeling of Nrf2 pathway activation in liver cells after toxicant exposure

Cells are exposed to oxidative stress and reactive metabolites every day. The Nrf2 signaling pathway responds to oxidative stress by upregulation of antioxidants like glutathione (GSH) to compensate the stress insult and re-establish homeostasis. Although mechanisms describing the interaction between the key pathway constituents Nrf2, Keap1 and p62 are widely reviewed and discussed in literature, quantitative dynamic models bringing together these mechanisms with time-resolved data are limited. Here, we present an ordinary differential equation (ODE) based dynamic model to describe the dynamic response of Nrf2, Keap1, Srxn1 and GSH to oxidative stress caused by the soft-electrophile diethyl maleate (DEM). The time-resolved data obtained by single-cell confocal microscopy of green fluorescent protein (GFP) reporters and qPCR of the Nrf2 pathway components complemented with siRNA knock down experiments, is accurately described by the calibrated mathematical model. We show that the quantitative model can describe the activation of the Nrf2 pathway by compounds with a different mechanism of activation, including drugs which are known for their ability to cause drug induced liver-injury (DILI) i.e., diclofenac (DCF) and omeprazole (OMZ). Finally, we show that our model can reveal differences in the processes leading to altered activation dynamics amongst DILI inducing drugs.

Mini-batch optimization enables training of ODE models on large-scale datasets

Quantitative dynamic models are widely used to study cellular signal processing. A critical step in modelling is the estimation of unknown model parameters from experimental data. As model sizes and datasets are steadily growing, established parameter optimization approaches for mechanistic models become computationally extremely challenging. Mini-batch optimization methods, as employed in deep learning, have better scaling properties. In this work, we adapt, apply, and benchmark mini-batch optimization for ordinary differential equation (ODE) models, thereby establishing a direct link between dynamic modelling and machine learning. On our main application example, a large-scale model of cancer signaling, we benchmark mini-batch optimization against established methods, achieving better optimization results and reducing computation by more than an order of magnitude. We expect that our work will serve as a first step towards mini-batch optimization tailored to ODE models and enable modelling of even larger and more complex systems than what is currently possible.

Boolean modelling in plant biology

Signalling and genetic networks underlie most biological processes and are often complex, containing many highly connected components. Modelling these networks can provide insight into mechanisms but is challenging given that rate parameters are often not well defined. Boolean modelling, in which components can only take on a binary value with connections encoded by logic equations, is able to circumvent some of these challenges, and has emerged as a viable tool to probe these complex networks. In this review, we will give an overview of Boolean modelling, with a specific emphasis on its use in plant biology. We review how Boolean modelling can be used to describe biological networks and then discuss examples of its applications in plant genetics and plant signalling.

Complementary resource preferences spontaneously emerge in diauxic microbial communities

Many microbes grow diauxically, utilizing the available resources one at a time rather than simultaneously. The properties of communities of microbes growing diauxically remain poorly understood, largely due to a lack of theory and models of such communities. Here, we develop and study a minimal model of diauxic microbial communities assembling in a serially diluted culture. We find that unlike co-utilizing communities, diauxic community assembly repeatably and spontaneously leads to communities with complementary resource preferences, namely communities where species prefer different resources as their top choice. Simulations and theory explain that the emergence of complementarity is driven by the disproportionate contribution of the top choice resource to the growth of a diauxic species. Additionally, we develop a geometric approach for analyzing serially diluted communities, with or without diauxie, which intuitively explains several additional emergent community properties, such as the apparent lack of species which grow fastest on a resource other than their most preferred resource. Overall, our work provides testable predictions for the assembly of natural as well as synthetic communities of diauxically shifting microbes.

Temporal assortment of cooperators in the spatial prisoner's dilemma

We study a spatial, one-shot prisoner's dilemma (PD) model in which selection operates on both an organism's behavioral strategy (cooperate or defect) and its decision of when to implement that strategy, which we depict as an organism's choice of one point in time, out of a set of discrete time slots, at which to carry out its PD strategy. Results indicate selection for cooperators across various time slots and parameter settings, including parameter settings in which cooperation would not evolve in an exclusively spatial model-as in work investigating exogenously imposed temporal networks. Moreover, in the presence of time slots, cooperators' portion of the population grows even under different combinations of spatial structure, transition rules, and update dynamics, though rates of cooperator fixation decline under pairwise comparison and synchronous updating. These findings indicate that, under certain evolutionary processes, merely existing in time and space promotes the evolution of cooperation.

Using predictive machine learning models for drug response simulation by calibrating patient-specific pathway signatures

The utility of pathway signatures lies in their capability to determine whether a specific pathway or biological process is dysregulated in a given patient. These signatures have been widely used in machine learning (ML) methods for a variety of applications including precision medicine, drug repurposing, and drug discovery. In this work, we leverage highly predictive ML models for drug response simulation in individual patients by calibrating the pathway activity scores of disease samples. Using these ML models and an intuitive scoring algorithm to modify the signatures of patients, we evaluate whether a given sample that was formerly classified as diseased, could be predicted as normal following drug treatment simulation. We then use this technique as a proxy for the identification of potential drug candidates. Furthermore, we demonstrate the ability of our methodology to successfully identify approved and clinically investigated drugs for four different cancers, outperforming six comparable state-of-the-art methods. We also show how this approach can deconvolute a drugs' mechanism of action and propose combination therapies. Taken together, our methodology could be promising to support clinical decision-making in personalized medicine by simulating a drugs' effect on a given patient.

Olive leaf-derived PPAR agonist complex induces collagen IV synthesis in human skin models

Introduction

Peroxisome proliferator‐activated receptor (PPAR) agonists are known to modulate the synthesis of dermal lipids and proteins including collagens. Olive (Olea europaea) leaves have been reported to contain PPAR‐binding ligands. Collagen IV, a major dermal‐epidermal junction (DEJ) protein, degrades with both age and disease. Here, we report the formulation of a novel multi‐ligand complex, Linefade, and its effects on collagen IV synthesis.

Methods

Linefade prepared from the leaves of Olea europaea contains 2% w/w plant extract solids dissolved in a mixture of glyceryl monoricinoleate and dimethyl isosorbide. In silico docking was performed with PPAR‐α (PDB ID: 2P54). Linefade was evaluated for PPAR‐α‐dependent transcription in a luciferase reporter assay system. Cell viability and collagen IV levels in human dermal fibroblast cultures were measured using the MTT method and ELISA assay, respectively. Transcriptome analysis was conducted on a full‐thickness reconstituted human skin (EpiDermFT) model. Ex vivo cell viability and collagen IV immunostaining were performed on human skin explants.

Results

In silico docking model of the major constituents (oleanolic acid and glyceryl monoricinoleate) produced a co‐binding affinity of −6.7 Kcal/mole. Linefade significantly increased PPAR‐α transcriptional activity in CHO cells and collagen IV synthesis in adult human dermal fibroblasts. Transcriptome analysis revealed that 1% Linefade modulated the expression of 280 genes with some related to epidermal differentiation, DEJ, PPAR, Nrf2 and retinoid pathways. An ex vivo human explant study showed that 1% Linefade, delivered via a triglycerides excipient, increased collagen IV levels along the dermal–epidermal junction by 52%.

Conclusion

In silico modelling and in vitro and ex vivo analyses confirmed Linefade‐mediated activation of PPAR‐α and stimulation of collagen IV synthesis.

Humans use minimum cost movements in a whole-body task

Humans have elegant bodies that allow gymnastics, piano playing, and tool use, but understanding how they do this in detail is difficult because their musculoskeletal systems are extraordinarily complicated. Nonetheless, common movements like walking and reaching can be stereotypical, and a very large number of studies have shown their energetic cost to be a major factor. In contrast, one might think that general movements are very individuated and intractable, but our previous study has shown that in an arbitrary set of whole-body movements used to trace large-scale closed curves, near-identical posture sequences were chosen across different subjects, both in the average trajectories of the body's limbs and in the variance within trajectories. The commonalities in that result motivate explanations for its generality. One explanation could be that humans also choose trajectories that are economical in cost. To test this hypothesis, we situate the tracing data within a forty eight degree of freedom human dynamic model that allows the computation of movement cost. Using the model to compare movement cost data from nominal tracings against various perturbed tracings shows that the latter are more energetically expensive, inferring that the original traces were chosen on the basis of minimum cost.

Commentary on Robinson et al. (2021): Evaluating theories of change for public health policies using computer model discovery methods

Recent developments in computer modelling—known as model discovery—could help to confirm the mechanisms underpinning Robinson and colleagues’ important early findings for the effectiveness of minimum unit pricing, and to test the complete theory of change underpinning this crucial evaluation.

Design principles of collateral sensitivity-based dosing strategies

Collateral sensitivity (CS)-based antibiotic treatments, where increased resistance to one antibiotic leads to increased sensitivity to a second antibiotic, may have the potential to limit the emergence of antimicrobial resistance. However, it remains unclear how to best design CS-based treatment schedules. To address this problem, we use mathematical modelling to study the effects of pathogen- and drug-specific characteristics for different treatment designs on bacterial population dynamics and resistance evolution. We confirm that simultaneous and one-day cycling treatments could supress resistance in the presence of CS. We show that the efficacy of CS-based cycling therapies depends critically on the order of drug administration. Finally, we find that reciprocal CS is not essential to suppress resistance, a result that significantly broadens treatment options given the ubiquity of one-way CS in pathogens. Overall, our analyses identify key design principles of CS-based treatment strategies and provide guidance to develop treatment schedules to suppress resistance.

Localization of Na+ channel clusters in narrowed perinexi of gap junctions enhances cardiac impulse transmission via ephaptic coupling: a model study

Abstract

It has been proposed that when gap junctional coupling is reduced in cardiac tissue, action potential propagation can be supported via ephaptic coupling, a mechanism mediated by negative electric potentials occurring in narrow intercellular clefts of intercalated discs (IDs). Recent studies showed that sodium (Na⁺) channels form clusters near gap junction plaques in nanodomains called perinexi, where the ID cleft is even narrower. To examine the electrophysiological relevance of Na⁺ channel clusters being located in perinexi, we developed a 3D finite element model of two longitudinally abutting cardiomyocytes, with a central Na⁺ channel cluster on the ID membranes. When this cluster was located in the perinexus of a closely positioned gap junction plaque, varying perinexal width greatly modulated impulse transmission from one cell to the other, with narrow perinexi potentiating ephaptic coupling. This modulation occurred via the interplay of Na⁺ currents, extracellular potentials in the cleft and patterns of current flow within the cleft. In contrast, when the Na⁺ channel cluster was located remotely from the gap junction plaque, this modulation by perinexus width largely disappeared. Interestingly, the Na⁺ current in the ID membrane of the pre‐junctional cell switched from inward to outward during excitation, thus contributing ions to the activating channels on the post‐junctional ID membrane. In conclusion, these results indicate that the localization of Na⁺ channel clusters in the perinexi of gap junction plaques is crucial for ephaptic coupling, which is furthermore greatly modulated by perinexal width. These findings are relevant for a comprehensive understanding of cardiac excitation.

Key points

Ephaptic coupling is a cardiac conduction mechanism involving nanoscale‐level interactions between the sodium (Na⁺) current and the extracellular potential in narrow intercalated disc clefts.

When gap junctional coupling is reduced, ephaptic coupling acts in conjunction with the classical cardiac conduction mechanism based on gap junctional current flow.

In intercalated discs, Na⁺ channels form clusters that are preferentially located in the periphery of gap junction plaques, in nanodomains known as perinexi, but the electrophysiological role of these perinexi has never been examined.

In our new 3D finite element model of two cardiac cells abutting each other with their intercalated discs, a Na⁺ channel cluster located inside a narrowed perinexus facilitated impulse transmission via ephaptic coupling.

Our simulations demonstrate the role of narrowed perinexi as privileged sites for ephaptic coupling in pathological situations when gap junctional coupling is decreased.

Development and qualification of a machine learning algorithm for automated hair counting

OBJECTIVE

Determining the amount of hair on the scalp has always been an important metric of patient satisfaction for hair growth and hair retention technologies. While simple in concept, this measurement is a difficult, resource intensive task for the dermatologist and the research scientist. Specifically, counting and measuring hair in phototrichogram images is very time consuming and labour intensive. Due to cost, often only a fraction of available images is manually analysed. There is a need for an automated method that can significantly increase speed and throughput while reducing the cost of counting and measuring hair in phototrichogram images.

METHODS

Recent advances in machine learning and deep convolutional neural networks (deep learning) have led to a revolution in the analysis of image, video, speech, text and other sensor data. Image diagnostics have seen remarkable improvements with completely automated methods outperforming both human experts and human-engineered analysis methods. Deep learning methods can also provide speed and cost benefits. To enable use of a deep learning, we created a data set of 288 manually annotated phototrichogram images with marked location and length of each hair (the training dataset). We designed a custom neural network architecture and custom image processing algorithms to best utilize the available training data and to maximize performance for hair counting and length measurement. The performance of the algorithm was qualified by comparing hair count and length measurements to an independent ground truth method, the semi-manual Canfield's Hair Metrix method.

RESULTS

Leveraging deep neural networks, we have developed capability to apply machine learning to reduce the time needed to acquire data from phototrichograms of patients' scalp from months to seconds. Our algorithm enables fast and fully automated hair counting and length measurement. The algorithm shows high agreement with human manually assisted analysis (ground truth).

CONCLUSIONS

We have trained and deployed an algorithm utilizing this technology and have demonstrated the reproducibility, accuracy and speed of this algorithm that, once deployed, requires little to no recurring cost or manual intervention for its operation. The method allows fast analysis of large number of images, reducing study cost and significantly reducing study analysis time.

Risk of proximal femoral nail antirotation (PFNA) implant failure upon different lateral femoral wall thickness in intertrochanteric fracture: a finite element analysis

Proximal Femoral Nail Antirotation (PFNA) has been commonly used to treat intertrochanteric fractures, despite the risk of implant failure. The integrity of the femur could influence the risk of implant failure. This study evaluated the influence of lateral femoral wall thickness on the potential of implant failure. A finite element model of the hip was reconstructed from the Computed Tomography of a female patient. Five intertrochanteric fracture models at different lateral femoral wall thickness (T1 = 27.6 mm, T2 = 25.4 mm, T3 = 23.4 mm, T4 = 21.4 mm, and T5 = 19.3 mm) were then created and fixed with PFNA. We simulated a critical loading condition by a high loading case during walking. Elastoplastic material models with yield stress and failure strain were applied to the bone and implant in which breakage can be simulated using the element deletion function. In addition, the stress and displacement of the implant and femur were analysed. Implant breakage occurred at the sides of the proximal nail canal in cases of T4 and T5 which was further supported by the higher maximum von Mises stress and nail displacement. The increased stress and displacement of the implant may implicate a reduction of stability and risk of implant failure. We suggested that precaution shall be taken when the wall thickness was less than 21.4 mm.