The benefits and pitfalls of machine learning for biomarker discovery

Development of an equation to predict delta bilirubin levels using machine learning

OBJECTIVE

Delta bilirubin (albumin-covalently bound bilirubin) may provide important clinical utility in identifying impaired hepatic excretion of conjugated bilirubin, but it cannot be measured in real-time for diagnostic purposes in clinical laboratories.

METHODS

A total of 210 samples were collected, and their delta bilirubin levels were measured four times using high-performance liquid chromatography. Data collected included age, sex, diagnosis code, delta bilirubin, total bilirubin, direct bilirubin, total protein, albumin, globulin, aspartate aminotransferase, alanine transaminase, alkaline phosphatase, gamma-glutamyl transferase, lactate dehydrogenase, hemoglobin, serum hemolysis value, hemolysis index, icterus value (Iv), icterus index (Ii), lipemia value (Lv), and lipemia index. To conduct feature selection and identify the optimal combination of variables, linear regression machine learning was performed 1,000 times.

RESULTS

The selected variables were total bilirubin, direct bilirubin, total protein, albumin, hemoglobin, Iv, Ii, and Lv. The best predictive performance for high delta bilirubin concentrations was achieved with the combination of albumin-direct bilirubin-hemoglobin-Iv-Lv. The final equation composed of these variables was as follows: delta bilirubin = 0.35 × Iv + 0.05 × Lv - 0.23 × direct bilirubin - 0.05 × hemoglobin - 0.04 × albumin + 0.10.

CONCLUSION

The equation established in this study is practical and can be easily applied in real-time in clinical laboratories.

Thrombosis, Translational Medicine, and Biomarker Research: Moving the Needle

Arterial and venous thromboembolism are leading causes of morbidity and death worldwide. Despite significant advances in the diagnosis, prognostication, and treatment of thrombotic diseases over the past 3 decades, the adoption of findings stemming from translational biomarker research in clinical practice remains limited. Biomarkers provide an opportunity to enhance our understanding of pathophysiological processes and optimize treatment strategies. They hold the promise of revolutionizing patient care. Still, this potential remains untapped, and several factors impede their use for near-patient applications. We sought to provide an overview of biomarker research in arterial and venous thromboembolic disease. We then aimed to discuss key barriers to the broader clinical implementation of biomarker research and highlight promising strategies to overcome them. We emphasize the merits of translational and implementation science to bridge the gaps from bench to bedside. Innovative trial design, data sharing, and collaborative efforts between academia and industry will be essential. Purposeful regression methodology using rational conceptual framework design, causal mediation analysis, and artificial intelligence might better leverage the use of observational data. Dedicated translational science training programs geared toward educating physicians on the appropriate measurement, interpretation, and integration of biomarker data in clinical practice should foster endorsement by frontline physicians. Finally, we make the case in support of a paradigm shift in cardiovascular medicine. Improved recognition of biomarker research and a greater emphasis on mechanistic evidence can better equip clinicians to deal with the uncertainty that defines the practice of thrombosis medicine.

Advancing miRNA cancer research through artificial intelligence: from biomarker discovery to therapeutic targeting

MicroRNAs (miRNAs), a class of small non-coding RNAs, play a vital role in regulating gene expression at the post-transcriptional level. Their discovery has profoundly impacted therapeutic strategies, particularly in cancer treatment, where RNA therapeutics, including miRNA-based targeted therapies, have gained prominence. Advances in RNA sequencing technologies have facilitated a comprehensive exploration of miRNAs-from fundamental research to their diagnostic and prognostic potential in various diseases, notably cancers. However, the manual handling and interpretation of vast RNA datasets pose significant challenges. The advent of artificial intelligence (AI) has revolutionized biological research by efficiently extracting insights from complex data. Machine learning algorithms, particularly deep learning techniques are effective for identifying critical miRNAs across different cancers and developing prognostic models. Moreover, the integration of AI has led to the creation of comprehensive miRNA databases for identifying mRNA and gene targets, thus facilitating deeper understanding and application in cancer research. This review comprehensively examines current developments in the application of machine learning techniques in miRNA research across diverse cancers. We discuss their roles in identifying biomarkers, elucidating miRNA targets, establishing disease associations, predicting prognostic outcomes, and exploring broader AI applications in cancer research. This review aims to guide researchers in leveraging AI techniques effectively within the miRNA field, thereby accelerating advancements in cancer diagnostics and therapeutics.

An exploration into the diagnostic capabilities of microRNAs for myocardial infarction using machine learning

BACKGROUND

MicroRNAs (miRNAs) have shown potential as diagnostic biomarkers for myocardial infarction (MI) due to their early dysregulation and stability in circulation after MI. Moreover, they play a crucial role in regulating adaptive and maladaptive responses in cardiovascular diseases, making them attractive targets for potential biomarkers. However, their potential as novel biomarkers for diagnosing cardiovascular diseases requires systematic evaluation.

METHODS

This study aimed to identify a miRNA biomarker panel for early-stage MI detection using bioinformatics and machine learning (ML) methods. miRNA expression data were obtained for early-stage MI patients and healthy controls from the Gene Expression Omnibus. Separate datasets were allocated for training and independent testing. Differential expression analysis was performed to identify dysregulated miRNAs in the training set. The least absolute shrinkage and selection operator (LASSO) was applied for feature selection to prioritize relevant miRNAs associated with MI. The selected miRNAs were used to develop ML models including support vector machine, Gradient Boosted, XGBoost, and a hard voting ensemble (HVE).

RESULTS

Differential expression analysis discovered 99 dysregulated miRNAs in the training set. LASSO feature selection prioritized 21 miRNAs. Ten miRNAs were identified in both the LASSO subset and independent test set. The HVE model trained with the selected miRNAs achieved an accuracy of 0.86 and AUC of 0.83 on the independent test set.

CONCLUSIONS

An integrated framework for robust miRNA selection from omics data shows promise for developing accurate diagnostic models for early-stage MI detection. The HVE model demonstrated good performance despite differences between training and test datasets.

The role of artificial intelligence in the management of liver diseases

Universal neonatal hepatitis B virus (HBV) vaccination and the advent of direct-acting antivirals (DAA) against hepatitis C virus (HCV) have reshaped the epidemiology of chronic liver diseases. However, some aspects of the management of chronic liver diseases remain unresolved. Nucleotide analogs can achieve sustained HBV DNA suppression but rarely lead to a functional cure. Despite the high efficacy of DAAs, successful antiviral therapy does not eliminate the risk of hepatocellular carcinoma (HCC), highlighted the need for cost-effective identification of high-risk populations for HCC surveillance and tailored HCC treatment strategies for these populations. The accessibility of high-throughput genomic data has accelerated the development of precision medicine, and the emergence of artificial intelligence (AI) has led to a new era of precision medicine. AI can learn from complex, non-linear data and identify hidden patterns within real-world datasets. The combination of AI and multi-omics approaches can facilitate disease diagnosis, biomarker discovery, and the prediction of treatment efficacy and prognosis. AI algorithms have been implemented in various aspects, including non-invasive tests, predictive models, image diagnosis, and the interpretation of histopathology findings. AI can support clinicians in decision-making, alleviate clinical burdens, and curtail healthcare expenses. In this review, we introduce the fundamental concepts of machine learning and review the role of AI in the management of chronic liver diseases.

Biomarkers From Discovery to Clinical Application: In Silico Pre-Clinical Validation Approach in the Face of Lung Cancer

Background

Clinical biomarkers, allow better classification of patients according to their disease risk, prognosis, and/or response to treatment. Although affordable omics-based approaches have paved the way for quicker identification of putative biomarkers, validation of biomarkers is necessary for translation of discoveries into clinical application.

Objective

Accordingly, in this study, we emphasize the potential of in silico approaches and have proposed and applied 3 novel sequential in silico pre-clinical validation steps to better identify the biomarkers that are truly desirable for clinical investment.

Design

As protein biomarkers are becoming increasingly important in the clinic alongside other molecular biomarkers and lung cancer is the most common cause of cancer-related deaths, we used protein biomarkers for lung cancer as an illustrative example to apply our in silico pre-clinical validation approach.

Methods

We collected the reported protein biomarkers for 3 cases (lung adenocarcinoma-LUAD, squamous cell carcinoma-LUSC, and unspecified lung cancer) and evaluated whether the protein biomarkers have cancer altering properties (i.e., act as tumor suppressors or oncoproteins and represent cancer hallmarks), are expressed in body fluids, and can be targeted by FDA-approved drugs.

Results

We collected 3008 protein biomarkers for lung cancer, 1189 for LUAD, and 182 for LUSC. Of these protein biomarkers for lung cancer, LUAD, and LUSC, only 28, 25, and 6 protein biomarkers passed the 3 in silico pre-clinical validation steps examined, and of these, only 5 and 2 biomarkers were specific for lung cancer and LUAD, respectively.

Conclusion

In this study, we applied our in silico pre-clinical validation approach the protein biomarkers for lung cancer cases. However, this approach can be applied and adapted to all cancer biomarkers. We believe that this approach will greatly facilitate the transition of cancer biomarkers into the clinical phase and offers great potential for future biomarker research.

Artificial Intelligence in Head and Neck Cancer: Innovations, Applications, and Future Directions

Artificial intelligence (AI) is revolutionizing head and neck cancer (HNC) care by providing innovative tools that enhance diagnostic accuracy and personalize treatment strategies. This review highlights the advancements in AI technologies, including deep learning and natural language processing, and their applications in HNC. The integration of AI with imaging techniques, genomics, and electronic health records is explored, emphasizing its role in early detection, biomarker discovery, and treatment planning. Despite noticeable progress, challenges such as data quality, algorithmic bias, and the need for interdisciplinary collaboration remain. Emerging innovations like explainable AI, AI-powered robotics, and real-time monitoring systems are poised to further advance the field. Addressing these challenges and fostering collaboration among AI experts, clinicians, and researchers is crucial for developing equitable and effective AI applications. The future of AI in HNC holds significant promise, offering potential breakthroughs in diagnostics, personalized therapies, and improved patient outcomes.

Progress in toxicogenomics to protect human health

Toxicogenomics measures molecular features, such as transcripts, proteins, metabolites and epigenomic modifications, to understand and predict the toxicological effects of environmental and pharmaceutical exposures. Transcriptomics has become an integral tool in contemporary toxicology research owing to innovations in gene expression profiling that can provide mechanistic and quantitative information at scale. These data can be used to predict toxicological hazards through the use of transcriptomic biomarkers, network inference analyses, pattern-matching approaches and artificial intelligence. Furthermore, emerging approaches, such as high-throughput dose-response modelling, can leverage toxicogenomic data for human health protection even in the absence of predicting specific hazards. Finally, single-cell transcriptomics and multi-omics provide detailed insights into toxicological mechanisms. Here, we review the progress since the inception of toxicogenomics in applying transcriptomics towards toxicology testing and highlight advances that are transforming risk assessment.

Revolutionary Point-of-Care Wearable Diagnostics for Early Disease Detection and Biomarker Discovery through Intelligent Technologies

Early-stage disease detection, particularly in Point-Of-Care (POC) wearable formats, assumes pivotal role in advancing healthcare services and precision-medicine. Public benefits of early detection extend beyond cost-effectively promoting healthcare outcomes, to also include reducing the risk of comorbid diseases. Technological advancements enabling POC biomarker recognition empower discovery of new markers for various health conditions. Integration of POC wearables for biomarker detection with intelligent frameworks represents ground-breaking innovations enabling automation of operations, conducting advanced large-scale data analysis, generating predictive models, and facilitating remote and guided clinical decision-making. These advancements substantially alleviate socioeconomic burdens, creating a paradigm shift in diagnostics, and revolutionizing medical assessments and technology development. This review explores critical topics and recent progress in development of 1) POC systems and wearable solutions for early disease detection and physiological monitoring, as well as 2) discussing current trends in adoption of smart technologies within clinical settings and in developing biological assays, and ultimately 3) exploring utilities of POC systems and smart platforms for biomarker discovery. Additionally, the review explores technology translation from research labs to broader applications. It also addresses associated risks, biases, and challenges of widespread Artificial Intelligence (AI) integration in diagnostics systems, while systematically outlining potential prospects, current challenges, and opportunities.

Tick exposure biomarkers: A One Health approach to new tick surveillance tools

The spread of tick-borne disease (TBD) is escalating globally, driven by climate change and socio-economic shifts, underlining the urgency to improve surveillance, diagnostics, and control strategies. Ticks can transmit a range of pathogens increasing the risk of transmission of human and veterinary diseases such as Lyme disease, tick-borne encephalitis, theileriosis, anaplasmosis, or Crimean-Congo hemorrhagic fever. Surveillance methods play a crucial role in monitoring the spread of tick-borne pathogens (TBP). However, there are shortcomings in the current surveillance methods regarding risks related to ticks. Human-tick encounters offer a novel metric for disease risk assessment, integrating human behavior into traditional surveillance models. However, to more reliably measure tick exposure, a molecular marker is needed. The identification of antibodies against arthropod salivary proteins as biomarkers for vector exposure represents a promising avenue for enhancing existing diagnostic and surveillance metrics. Here we explore how the use of tick saliva biomarkers targeting recombinant proteins and synthetic peptides could significantly improve the assessment of TBD transmission risk and the effectiveness of vector control measures. With focused efforts on creating a biomarker against tick exposure suitable for humans and domestic animals alike, tick surveillance, diagnosis and control would be more achievable and aid in reducing the mounting threat of TBP through a One Health lens.

SuperAging functional connectomics from resting-state functional MRI

Understanding the relationship between functional connectivity (FC) of higher-order neurocognitive networks and age-related cognitive decline is a complex and evolving field of research. Decreases in FC have been associated with cognitive decline in persons with Alzheimer's disease and related dementias (ADRD). However, the contributions of FC have been less straightforward in typical cognitive aging. Some investigations suggest relatively robust FC within neurocognitive networks differentiates unusually successful cognitive aging from average aging, while others do not. Methodologic limitations in data processing and varying definitions of 'successful aging' may have contributed to the inconsistent results to date. The current study seeks to address previous limitations by optimized MRI methods to examine FC in the well-established SuperAging phenotype, defined by age and cognitive performance as individuals 80 and older with episodic memory performance equal to or better than 50-to-60-year-olds. Within- and between-network FC of large-scale neurocognitive networks were compared between 24 SuperAgers and 16 cognitively average older-aged control (OACs) with stable cognitive profiles using resting-state functional MRI (rs-fMRI) from a single visit. Group classification was determined based on measures of episodic memory, executive functioning, verbal fluency and picture naming. Inclusion criteria required stable cognitive status across two visits. First, we investigated the FC within and between seven resting-state networks from a common atlas parcellation. A separate index of network segregation was also compared between groups. Second, we investigated the FC between six subcomponents of the default mode network (DMN), the neurocognitive network commonly associated with memory performance and disrupted in persons with ADRD. For each analysis, FCs were compared across groups using two-sample independent t-tests and corrected for multiple comparisons. There were no significant between-group differences in demographic characteristics including age, sex and education. At the group-level, within-network FC, between-network FC, and segregation measurements of seven large-scale networks, including subcomponents of the DMN, were not a primary differentiator between cognitively average aging and SuperAging phenotypes. Thus, FC within or between large-scale networks does not appear to be a primary driver of the exceptional memory performance observed in SuperAgers. These results have relevance for differentiating the role of FC changes associated with cognitive aging from those associated with ADRD.

Bioinformatics and machine learning to support nanomaterial grouping

Nanomaterials (NMs) offer plenty of novel functionalities. Moreover, their physicochemical properties can be fine-tuned to meet the needs of specific applications, leading to virtually unlimited numbers of NM variants. Hence, efficient hazard and risk assessment strategies building on New Approach Methodologies (NAMs) become indispensable. Indeed, the design, the development and implementation of NAMs has been a major topic in a substantial number of research projects. One of the promising strategies that can help to deal with the high number of NMs variants is grouping and read-across. Based on demonstrated structural and physicochemical similarity, NMs can be grouped and assessed together. Within an established NM group, read-across may be performed to fill in data gaps for data-poor variants using existing data for NMs within the group. Establishing a group requires a sound justification, usually based on a grouping hypothesis that links specific physicochemical properties to well-defined hazard endpoints. However, for NMs these interrelationships are only beginning to be understood. The aim of this review is to demonstrate the power of bioinformatics with a specific focus on Machine Learning (ML) approaches to unravel the NM Modes-of-Action (MoA) and identify the properties that are relevant to specific hazards, in support of grouping strategies. This review emphasizes the following messages: 1) ML supports identification of the most relevant properties contributing to specific hazards; 2) ML supports analysis of large omics datasets and identification of MoA patterns in support of hypothesis formulation in grouping approaches; 3) omics approaches are useful for shifting away from consideration of single endpoints towards a more mechanistic understanding across multiple endpoints gained from one experiment; and 4) approaches from other fields of Artificial Intelligence (AI) like Natural Language Processing or image analysis may support automated extraction and interlinkage of information related to NM toxicity. Here, existing ML models for predicting NM toxicity and for analyzing omics data in support of NM grouping are reviewed. Various challenges related to building robust models in the field of nanotoxicology exist and are also discussed.

Development of a long noncoding RNA-based machine learning model to predict COVID-19 in-hospital mortality

Tools for predicting COVID-19 outcomes enable personalized healthcare, potentially easing the disease burden. This collaborative study by 15 institutions across Europe aimed to develop a machine learning model for predicting the risk of in-hospital mortality post-SARS-CoV-2 infection. Blood samples and clinical data from 1286 COVID-19 patients collected from 2020 to 2023 across four cohorts in Europe and Canada were analyzed, with 2906 long non-coding RNAs profiled using targeted sequencing. From a discovery cohort combining three European cohorts and 804 patients, age and the long non-coding RNA LEF1-AS1 were identified as predictive features, yielding an AUC of 0.83 (95% CI 0.82-0.84) and a balanced accuracy of 0.78 (95% CI 0.77-0.79) with a feedforward neural network classifier. Validation in an independent Canadian cohort of 482 patients showed consistent performance. Cox regression analysis indicated that higher levels of LEF1-AS1 correlated with reduced mortality risk (age-adjusted hazard ratio 0.54, 95% CI 0.40-0.74). Quantitative PCR validated LEF1-AS1's adaptability to be measured in hospital settings. Here, we demonstrate a promising predictive model for enhancing COVID-19 patient management.

Multicomponent (bio)markers for obesity risk prediction: a scoping review protocol

INTRODUCTION

Despite international efforts, the number of individuals struggling with obesity is still increasing. An important aspect of obesity prevention relates to identifying individuals at risk at early stage, allowing for timely risk stratification and initiation of countermeasures. However, obesity is complex and multifactorial by nature, and one isolated (bio)marker is unlikely to enable an optimal risk stratification and prognosis for the individual; rather, a combined set is required. Such a multicomponent interpretation would integrate biomarkers from various domains, such as classical markers (eg, anthropometrics, blood lipids), multiomics (eg, genetics, proteomics, metabolomics), lifestyle and behavioural attributes (eg, diet, physical activity, sleep patterns), psychological traits (mental health status such as depression) and additional host factors (eg, gut microbiota diversity), also by means of advanced interpretation tools such as machine learning. In this paper, we will present a protocol that will be employed for a scoping review that attempts to summarise and map the state-of-the-art in the area of multicomponent (bio)markers related to obesity, focusing on the usability and effectiveness of such biomarkers.

METHODS AND ANALYSIS

PubMed, Scopus, CINAHL and Embase databases will be searched using predefined key terms to identify peer-reviewed articles published in English until January 2024. Once downloaded into EndNote for deduplication, CADIMA will be employed to review and select abstracts and full-text articles in a two-step procedure, by two independent reviewers. Data extraction will then be carried out by several independent reviewers. Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews and Peer Review of Electronic Search Strategies guidelines will be followed. Combinations employing at least two biomarkers from different domains will be mapped and discussed.

ETHICS AND DISSEMINATION

Ethical approval is not required; data will rely on published articles. Findings will be published open access in an international peer-reviewed journal. This review will allow guiding future directions for research and public health strategies on obesity prevention, paving the way towards multicomponent interventions.

Development and validation of a cuproptosis-related prognostic model for acute myeloid leukemia patients using machine learning with stacking

Our objective is to develop a prognostic model focused on cuproptosis, aimed at predicting overall survival (OS) outcomes among Acute myeloid leukemia (AML) patients. The model utilized machine learning algorithms incorporating stacking. The GSE37642 dataset was used as the training data, and the GSE12417 and TCGA-LAML cohorts were used as the validation data. Stacking was used to merge the three prediction models, subsequently using a random survival forests algorithm to refit the final model using the stacking linear predictor and clinical factors. The prediction model, featuring stacking linear predictor and clinical factors, achieved AUC values of 0.840, 0.876 and 0.892 at 1, 2 and 3 years within the GSE37642 dataset. In external validation dataset, the corresponding AUCs were 0.741, 0.754 and 0.783. The predictive performance of the model in the external dataset surpasses that of the model simply incorporates all predictors. Additionally, the final model exhibited good calibration accuracy. In conclusion, our findings indicate that the novel prediction model refines the prognostic prediction for AML patients, while the stacking strategy displays potential for model integration.

A 19-Gene Signature of Serous Ovarian Cancer Identified by Machine Learning and Systems Biology: Prospects for Diagnostics and Personalized Medicine

Ovarian cancer is a major cause of cancer deaths among women. Early diagnosis and precision/personalized medicine are essential to reduce mortality and morbidity of ovarian cancer, as with new molecular targets to accelerate drug discovery. We report here an integrated systems biology and machine learning (ML) approach based on the differential coexpression analysis to identify candidate systems biomarkers (i.e., gene modules) for serous ovarian cancer. Accordingly, four independent transcriptome datasets were statistically analyzed independently and common differentially expressed genes (DEGs) were identified. Using these DEGs, coexpressed gene pairs were unraveled. Subsequently, differential coexpression networks between the coexpressed gene pairs were reconstructed so as to identify the differentially coexpressed gene modules. Based on the established criteria, "SOV-module" was identified as being significant, consisting of 19 genes. Using independent datasets, the diagnostic capacity of the SOV-module was evaluated using principal component analysis (PCA) and ML techniques. PCA showed a sensitivity and specificity of 96.7% and 100%, respectively, and ML analysis showed an accuracy of up to 100% in distinguishing phenotypes in the present study sample. The prognostic capacity of the SOV-module was evaluated using survival and ML analyses. We found that the SOV-module's performance for prognostics was significant (p-value = 1.36 × 10-4) with an accuracy of 63% in discriminating between survival and death using ML techniques. In summary, the reported genomic systems biomarker candidate offers promise for personalized medicine in diagnosis and prognosis of serous ovarian cancer and warrants further experimental and translational clinical studies.