Targeted deep learning classification and feature extraction for clinical diagnosis.

Protein biomarkers can be used to characterize symptom classes, which describe the metabolic or immunodeficient state of patients during the progression of a specific disease. Recent literature has shown that machine learning methods can complement traditional clinical methods in identifying biomarkers. However, many machine learning frameworks only apply narrowly to a specific archetype or subset of diseases. In this paper, we propose a feature extractor which can discover protein biomarkers for a wide variety of classification problems. The feature extractor uses a special type of deep learning model, which discovers a latent space that allows for optimal class separation and enhanced class cluster identity. The extracted biomarkers can then be used to train highly accurate supervised learning models. We apply our methods to a dataset involving COVID-19 patients and another involving scleroderma patients, to demonstrate improved class separation and reduced false discovery rates compared to results obtained using traditional models.

Identifying indicator species in ecological habitats using Deep Optimal Feature Learning.

Much of the current research on supervised modelling is focused on maximizing outcome prediction accuracy. However, in engineering disciplines, an arguably more important goal is that of feature extraction, the identification of relevant features associated with the various outcomes. For instance, in microbial communities, the identification of keystone species can often lead to improved prediction of future behavioral shifts. This paper proposes a novel feature extractor based on Deep Learning, which is largely agnostic to underlying assumptions regarding the training data. Starting from a collection of microbial species abundance counts, the Deep Learning model first trains itself to classify the selected distinct habitats. It then identifies indicator species associated with the habitats. The results are then compared and contrasted with those obtained by traditional statistical techniques. The indicator species are similar when compared at top taxonomic levels such as Domain and Phylum, despite visible differences in lower levels such as Class and Order. More importantly, when our estimated indicators are used to predict final habitat labels using simpler models (such as Support Vector Machines and traditional Artificial Neural Networks), the prediction accuracy is improved. Overall, this study serves as a preliminary step that bridges modern, black-box Machine Learning models with traditional, domain expertise-rich techniques.

Effects of cysteine, asparagine, or glutamine limitations in Chinese hamster ovary cell batch and fed-batch cultures.

Amino acid availability is a key factor that can be controlled to optimize the productivity of fed-batch cultures. To study amino acid limitation effects, a serum-free chemically defined basal medium was formulated to exclude the amino acids that became depleted in batch culture. The effect of limiting glutamine, asparagine, and cysteine on the cell growth, metabolism, antibody productivity, and product glycosylation was investigated in three Chinese hamster ovary (CHO) cell lines (CHO-DXB11, CHO-K1SV, and CHO-S). Cysteine limitation was detrimental to both cell proliferation and productivity for all three CHO cell lines. Glutamine limitation reduced growth but not cell specific productivity, whereas asparagine limitation had no significant effect on either growth or cell specific productivity. Neither glutamine nor asparagine limitation significantly affected antibody glycosylation. Replenishing the CHO-DXB11 culture with cysteine after 1 day of cysteine limitation allowed the cells to partially recover their growth and productivity. This recovery was not observed after 2 days of cysteine limitation. Based on these findings, a fed-batch protocol was developed using single or mixed amino acid supplementation. Although cell density and antibody concentration were lower compared to a commercial feed, the feeds based on cysteine supplementation yielded comparable cell specific productivity. Overall, this study showed that different amino acid limitations have varied effects on the performance of CHO cell cultures and that maintaining cysteine availability is a critical process parameter for the three cell lines investigated.

Assessment of type II diabetes mellitus using irregularly sampled measurements with missing data.

Diabetes mellitus is one of the leading diseases in the developed world. In order to better regulate blood glucose in a diabetic patient, improved modelling of insulin-glucose dynamics is a key factor in the treatment of diabetes mellitus. In the current work, the insulin-glucose dynamics in type II diabetes mellitus can be modelled by using a stochastic nonlinear state-space model. Estimating the parameters of such a model is difficult as only a few blood glucose and insulin measurements per day are available in a non-clinical setting. Therefore, developing a predictive model of the blood glucose of a person with type II diabetes mellitus is important when the glucose and insulin concentrations are only available at irregular intervals. To overcome these difficulties, we resort to online sequential Monte Carlo (SMC) estimation of states and parameters of the state-space model for type II diabetic patients under various levels of randomly missing clinical data. Our results show that this method is efficient in monitoring and estimating the dynamics of the peripheral glucose, insulin and incretins concentration when 10, 25 and 50% of the simulated clinical data were randomly removed.

Use of Molecular Dynamics for the Refinement of an Electrostatic Model for the In Silico Design of a Polymer Antidote for the Anticoagulant Fondaparinux.

Molecular dynamics (MD) simulations results are herein incorporated into an electrostatic model used to determine the structure of an effective polymer-based antidote to the anticoagulant fondaparinux. In silico data for the polymer or its cationic binding groups has not, up to now, been available, and experimental data on the structure of the polymer-fondaparinux complex is extremely limited. Consequently, the task of optimizing the polymer structure is a daunting challenge. MD simulations provided a means to gain microscopic information on the interactions of the binding groups and fondaparinux that would have otherwise been inaccessible. This was used to refine the electrostatic model and improve the quantitative model predictions of binding affinity. Once refined, the model provided guidelines to improve electrostatic forces between candidate polymers and fondaparinux in order to increase association rate constants.

Pharmacokinetic-Pharmacodynamic Modeling of Metformin for the Treatment of Type II Diabetes Mellitus.

Metformin is an antihyperglycemic agent commonly used for the treatment of Type II diabetes mellitus. However, its effects on patients are derived usually from clinical experiments. In this study, a dynamic model of Type II diabetes mellitus with the treatment of metformin is proposed. The Type II diabetic model is a modification of an existing compartmental diabetic model. The dynamic simulation of the metformin effect for a Type II diabetic patient is based on the pharmacokinetic and pharmacodynamic relationship with a human body. The corresponding model parameters are estimated by optimization using clinical data from published reports. Then, the effect of metformin in both intravenous and oral administration on a Type II diabetes mellitus model are compared. The combination treatment of insulin infusion plus oral metformin is shown to be superior than the monotherapy with oral metformin only. These results are consistent with the clinical understanding of the use of metformin. For further work, the model can be analyzed for evaluating the treatment of diabetes mellitus with different pharmacological agents.