Global, local and personalised modeling and pattern discovery in bioinformatics: An integrated approach

Filter and Wrapper Stacking Ensemble (FWSE): a robust approach for reliable biomarker discovery in high-dimensional omics data

Selecting informative features, such as accurate biomarkers for disease diagnosis, prognosis and response to treatment, is an essential task in the field of bioinformatics. Medical data often contain thousands of features and identifying potential biomarkers is challenging due to small number of samples in the data, method dependence and non-reproducibility. This paper proposes a novel ensemble feature selection method, named Filter and Wrapper Stacking Ensemble (FWSE), to identify reproducible biomarkers from high-dimensional omics data. In FWSE, filter feature selection methods are run on numerous subsets of the data to eliminate irrelevant features, and then wrapper feature selection methods are applied to rank the top features. The method was validated on four high-dimensional medical datasets related to mental illnesses and cancer. The results indicate that the features selected by FWSE are stable and statistically more significant than the ones obtained by existing methods while also demonstrating biological relevance. Furthermore, FWSE is a generic method, applicable to various high-dimensional datasets in the fields of machine intelligence and bioinformatics.

Constrained neuro fuzzy inference methodology for explainable personalised modelling with applications on gene expression data

Interpretable machine learning models for gene expression datasets are important for understanding the decision-making process of a classifier and gaining insights on the underlying molecular processes of genetic conditions. Interpretable models can potentially support early diagnosis before full disease manifestation. This is particularly important yet, challenging for mental health. We hypothesise this is due to extreme heterogeneity issues which may be overcome and explained by personalised modelling techniques. Thus far, most machine learning methods applied to gene expression datasets, including deep neural networks, lack personalised interpretability. This paper proposes a new methodology named personalised constrained neuro fuzzy inference (PCNFI) for learning personalised rules from high dimensional datasets which are structurally and semantically interpretable. Case studies on two mental health related datasets (schizophrenia and bipolar disorders) have shown that the relatively short and simple personalised fuzzy rules provided enhanced interpretability as well as better classification performance compared to other commonly used machine learning methods. Performance test on a cancer dataset also showed that PCNFI matches previous benchmarks. Insights from our approach also indicated the importance of two genes (ATRX and TSPAN2) as possible biomarkers for early differentiation of ultra-high risk, bipolar and healthy individuals. These genes are linked to cognitive ability and impulsive behaviour. Our findings suggest a significant starting point for further research into the biological role of cognitive and impulsivity-related differences. With potential applications across bio-medical research, the proposed PCNFI method is promising for diagnosis, prognosis, and the design of personalised treatment plans for better outcomes in the future.

Development and Validation of a Personalized Model With Transfer Learning for Acute Kidney Injury Risk Estimation Using Electronic Health Records

IMPORTANCE

Acute kidney injury (AKI) is a heterogeneous syndrome prevalent among hospitalized patients. Personalized risk estimation and risk factor identification may allow effective intervention and improved outcomes.

OBJECTIVE

To develop and validate personalized AKI risk estimation models using electronic health records (EHRs), examine whether personalized models were beneficial in comparison with global and subgroup models, and assess the heterogeneity of risk factors and their outcomes in different subpopulations.

DESIGN, SETTING, AND PARTICIPANTS

This diagnostic study analyzed EHR data from 1 tertiary care hospital and used machine learning and logistic regression to develop and validate global, subgroup, and personalized risk estimation models. Transfer learning was implemented to enhance the personalized model. Predictor outcomes across subpopulations were analyzed, and metaregression was used to explore predictor interactions. Adults who were hospitalized for 2 or more days from November 1, 2007, to December 31, 2016, were included in the analysis. Patients with moderate or severe kidney dysfunction at admission were excluded. Data were analyzed between August 28, 2019, and May 8, 2022.

EXPOSURES

Clinical and laboratory variables in the EHR.

MAIN OUTCOMES AND MEASURES

The main outcome was AKI of any severity, and AKI was defined using the Kidney Disease: Improving Global Outcomes serum creatinine criteria. Performance of the models was measured with area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve, and calibration.

RESULTS

The study cohort comprised 76 957 inpatient encounters. Patients had a mean (SD) age of 55.5 (17.4) years and included 42 159 men (54.8%). The personalized model with transfer learning outperformed the global model for AKI estimation in terms of AUROC among general inpatients (0.78 [95% CI, 0.77-0.79] vs 0.76 [95% CI, 0.75-0.76]; P < .001) and across the high-risk subgroups (0.79 [95% CI, 0.78-0.80] vs 0.75 [95% CI, 0.74-0.77]; P < .001) and low-risk subgroups (0.74 [95% CI, 0.73-0.75] vs 0.71 [95% CI, 0.70-0.72]; P < .001). The AUROC improvement reached 0.13 for the high-risk subgroups, such as those undergoing liver transplant and cardiac surgery. Moreover, the personalized model with transfer learning performed better than or comparably with the best published models in well-studied AKI subgroups. Predictor outcomes varied significantly between patients, and interaction analysis uncovered modifiers of the predictor outcomes.

CONCLUSIONS AND RELEVANCE

Results of this study demonstrated that a personalized modeling with transfer learning is an improved AKI risk estimation approach that can be used across diverse patient subgroups. Risk factor heterogeneity and interactions at the individual level highlighted the need for agile, personalized care.

An Investigation of How Normalisation and Local Modelling Techniques Confound Machine Learning Performance In a Mental Health Study

Machine learning (ML) is increasingly deployed on biomedical studies for biomarker development (feature selection) and diagnostic/prognostic technologies (classification). While different ML techniques produce different feature sets and classification performances, less understood is how upstream data processing methods (e.g., normalisation) impact downstream analyses. Using a clinical mental health dataset, we investigated the impact of different normalisation techniques on classification model performance. Gene Fuzzy Scoring (GFS), an in-house developed normalisation technique, is compared against widely used normalisation methods such as global quantile normalisation, class-specific quantile normalisation and surrogate variable analysis. We report that choice of normalisation technique has strong influence on feature selection. with GFS outperforming other techniques. Although GFS parameters are tuneable, good classification model performance (ROC-AUC > 0.90) is observed regardless of the GFS parameter settings. We also contrasted our results against local modelling, which is meant to improve the resolution and meaningfulness of classification models built on heterogeneous data. Local models, when derived from non-biologically meaningful subpopulations, perform worse than global models. A deep dive however, revealed that the factors driving cluster formation has little to do with the phenotype-of-interest. This finding is critical, as local models are often seen as a superior means of clinical data modelling. We advise against such naivete. Additionally, we have developed a combinatorial reasoning approach using both global and local paradigms: This helped reveal potential data quality issues or underlying factors causing data heterogeneity that are often overlooked. It also assists to explain the model as well as provides directions for further improvement.

Proposing Two Local Modeling Approaches for Discriminating PGI Sunite Lamb from Other Origins Using Stable Isotopes and Machine Learning

For the protection of Protected Geographical Indication (PGI) Sunite lamb, PGI Sunite lamb samples and lamb samples from two other banners in the Inner Mongolia autonomous region were distinguished by stable isotopes (δ¹³C, δ¹⁵N, δ²H, and δ¹⁸O) and two local modeling approaches. In terms of the main characteristics and predictive performance, local modeling was better than global modeling. The accuracies of five local models (LDA, RF, SVM, BPNN, and KNN) obtained by the Adaptive Kennard–Stone algorithm were 91.30%, 95.65%, 91.30%, 100%, and 91.30%, respectively. The accuracies of the five local models obtained by an approach of PCA–Full distance based on DD–SIMCA were 91.30%, 91.30%, 91.30%, 100%, and 95.65%, respectively. The accuracies of the five global models were 91.30%, 91.30%, 91.30%, 100%, and 91.30%, respectively. Stable isotope ratio analysis combined with local modeling can be used as an effective indicator for protecting PGI Sunite lamb.

Personalised predictive modelling with brain-inspired spiking neural networks of longitudinal MRI neuroimaging data and the case study of dementia

BACKGROUND

Longitudinal neuroimaging provides spatiotemporal brain data (STBD) measurement that can be utilised to understand dynamic changes in brain structure and/or function underpinning cognitive activities. Making sense of such highly interactive information is challenging, given that the features manifest intricate temporal, causal relations between the spatially distributed neural sources in the brain.

METHODS

The current paper argues for the advancement of deep learning algorithms in brain-inspired spiking neural networks (SNN), capable of modelling structural data across time (longitudinal measurement) and space (anatomical components). The paper proposes a methodology and a computational architecture based on SNN for building personalised predictive models from longitudinal brain data to accurately detect, understand, and predict the dynamics of an individual's functional brain state. The methodology includes finding clusters of similar data to each individual, data interpolation, deep learning in a 3-dimensional brain-template structured SNN model, classification and prediction of individual outcome, visualisation of structural brain changes related to the predicted outcomes, interpretation of results, and individual and group predictive marker discovery.

RESULTS

To demonstrate the functionality of the proposed methodology, the paper presents experimental results on a longitudinal magnetic resonance imaging (MRI) dataset derived from 175 older adults of the internationally recognised community-based cohort Sydney Memory and Ageing Study (MAS) spanning 6 years of follow-up.

SIGNIFICANCE

The models were able to accurately classify and predict 2 years ahead of cognitive decline, such as mild cognitive impairment (MCI) and dementia with 95% and 91% accuracy, respectively. The proposed methodology also offers a 3-dimensional visualisation of the MRI models reflecting the dynamic patterns of regional changes in white matter hyperintensity (WMH) and brain volume over 6 years.

CONCLUSION

The method is efficient for personalised predictive modelling on a wide range of neuroimaging longitudinal data, including also demographic, genetic, and clinical data. As a case study, it resulted in finding predictive markers for MCI and dementia as dynamic brain patterns using MRI data.

Feature-based multi-criteria recommendation system using a weighted approach with ranking correlation

With the increase of online businesses, recommendation algorithms are being researched a lot to facilitate the process of using the existing information. Such multi-criteria recommendation (MCRS) helps a lot the end-users to attain the required results of interest having different selective criteria – such as combinations of implicit and explicit interest indicators in the form of ranking or rankings on different matched dimensions. Current approaches typically use label correlation, by assuming that the label correlations are shared by all objects. In real-world tasks, however, different sources of information have different features. Recommendation systems are more effective if being used for making a recommendation using multiple criteria of decisions by using the correlation between the features and items content (content-based approach) or finding a similar user rating to get targeted results (Collaborative filtering). To combine these two filterings in the multicriteria model, we proposed a features-based fb-knn multi-criteria hybrid recommendation algorithm approach for getting the recommendation of the items by using multicriteria features of items and integrating those with the correlated items found in similar datasets. Ranks were assigned to each decision and then weights were computed for each decision by using the standard deviation of items to get the nearest result. For evaluation, we tested the proposed algorithm on different datasets having multiple features of information. The results demonstrate that proposed fb-knn is efficient in different types of datasets.

Personalised modelling with spiking neural networks integrating temporal and static information

This paper proposes a new personalised prognostic/diagnostic system that supports classification, prediction and pattern recognition when both static and dynamic/spatiotemporal features are presented in a dataset. The system is based on a proposed clustering method (named d2WKNN) for optimal selection of neighbouring samples to an individual with respect to the integration of both static (vector-based) and temporal individual data. The most relevant samples to an individual are selected to train a Personalised Spiking Neural Network (PSNN) that learns from sets of streaming data to capture the space and time association patterns. The generated time-dependant patterns resulted in a higher accuracy of classification/prediction (80% to 93%) when compared with global modelling and conventional methods. In addition, the PSNN models can support interpretability by creating personalised profiling of an individual. This contributes to a better understanding of the interactions between features. Therefore, an end-user can comprehend what interactions in the model have led to a certain decision (outcome). The proposed PSNN model is an analytical tool, applicable to several real-life health applications, where different data domains describe a person's health condition. The system was applied to two case studies: (1) classification of spatiotemporal neuroimaging data for the investigation of individual response to treatment and (2) prediction of risk of stroke with respect to temporal environmental data. For both datasets, besides the temporal data, static health data were also available. The hyper-parameters of the proposed system, including the PSNN models and the d2WKNN clustering parameters, are optimised for each individual.

A new local-global approach for classification

In this paper, we propose a new local-global pattern classification scheme that combines supervised and unsupervised approaches, taking advantage of both, local and global environments. We understand as global methods the ones concerned with the aim of constructing a model for the whole problem space using the totality of the available observations. Local methods focus into sub regions of the space, possibly using an appropriately selected subset of the sample. In the proposed method, the sample is first divided in local cells by using a Vector Quantization unsupervised algorithm, the LBG (Linde-Buzo-Gray). In a second stage, the generated assemblage of much easier problems is locally solved with a scheme inspired by Bayes' rule. Four classification methods were implemented for comparison purposes with the proposed scheme: Learning Vector Quantization (LVQ); Feedforward Neural Networks; Support Vector Machine (SVM) and k-Nearest Neighbors. These four methods and the proposed scheme were implemented in eleven datasets, two controlled experiments, plus nine public available datasets from the UCI repository. The proposed method has shown a quite competitive performance when compared to these classical and largely used classifiers. Our method is simple concerning understanding and implementation and is based on very intuitive concepts.

Principles for 3D/4D QSAR classification of drugs

The principles for the 3D/4D classification of drugs are introduced in this article. Based on these principles, new techniques for the reconstruction of complementary selfconsistent receptor fields for the classification of drugs, taking into account their multitautomeric and multiconformational states, are created. The series of examples of classification of drugs by their activity (active or nonactive), by their mechanisms of action, by their target and binding site and by the most important stages of their action are given. Prospects for rational drug design are highlighted.

A new paradigm for pattern recognition of drugs

A new paradigm is suggested for pattern recognition of drugs. The approach is based on the combined application of the 4D/3D quantitative structure-activity relationship (QSAR) algorithms BiS and ConGO. The first algorithm, BiS/MC (multiconformational), is used for the search for the conformers interacting with a receptor. The second algorithm, ConGO, has been suggested for the detailed study of the selected conformers' electron density and for the search for the electron structure fragments that determine the pharmacophore and antipharmacophore parts of the compounds. In this work we suggest using a new AlteQ method for the evaluation of the molecular electron density. AlteQ describes the experimental electron density (determined by low-temperature highly accurate X-ray analysis) much better than a number of quantum approaches. Herein this is shown using a comparison of the computed electron density with the results of highly accurate X-ray analysis. In the present study the desirability function is used for the first time for the analysis of the effects of the electron structure in the process of pattern recognition of active and inactive compounds. The suggested method for pattern recognition has been used for the investigation of various sets of compounds such as DNA-antimetabolites, fXa inhibitors, 5-HT(1A), and alpha(1)-AR receptors inhibitors. The pharmacophore and antipharmacophore fragments have been found in the electron structures of the compounds. It has been shown that the pattern recognition cross-validation quality for the datasets is unity.