Multimodal AI/ML for discovering novel biomarkers and predicting disease using multi-omics profiles of patients with cardiovascular diseases

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Ноя. 3, 2024

Cardiovascular diseases (CVDs) are complex, multifactorial conditions that require personalized assessment and treatment. Advancements in multi-omics technologies, namely RNA sequencing whole-genome sequencing, have provided translational researchers with a comprehensive view of the human genome. The efficient synthesis analysis this data through integrated approach characterizes genetic variants alongside expression patterns linked to emerging phenotypes, can reveal novel biomarkers enable segmentation patient populations based on risk factors. In study, we present cutting-edge methodology rooted integration traditional bioinformatics, classical statistics, multimodal machine learning techniques. Our has potential uncover intricate mechanisms underlying CVD, enabling patient-specific response profiling. We sourced transcriptomic single nucleotide polymorphisms (SNPs) from both CVD patients healthy controls. By integrating these datasets clinical demographic information, generated profiles. Utilizing robust feature selection approach, identified signature 27 features SNPs effective predictors CVD. Differential analysis, combined minimum redundancy maximum relevance selection, highlighted explain disease phenotype. This prioritizes biological efficiency learning. employed Combination Annotation Dependent Depletion scores allele frequencies identify pathogenic characteristics patients. Classification models trained demonstrated high-accuracy predictions for best performing was an XGBoost classifier optimized via Bayesian hyperparameter tuning, which able correctly classify all our test dataset. Using SHapley Additive exPlanations, created assessments patients, offering further contextualization setting. Across cohort, RPL36AP37 HBA1 were scored as most important predicting CVDs. A literature review revealed substantial portion diagnostic previously been associated framework propose study is unbiased generalizable other disorders.

Язык: Английский

Association of cardiovascular disease on cancer: observational and mendelian randomization analyses

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Дек. 24, 2024

Extensive research is needed to examine the association between cardiovascular disease (CVD) and cancer. The observational study based on data collected from 2005–2018 National Health Nutrition Examination Survey (NHANES). To assess connection CVDs cancer, we used a weighted multivariable logistic regression analysis with as many confounding factors feasible included in model. By employing Mendelian randomization (MR), unbiased causal relationship cancers was ascertained. primary analytical approach employed Inverse Variance Weighted methodology. In cross-sectional study, positive correlation observed CVD cancer (Model 3, Odds ratio 1.26, 95% confidence interval 1.01 ~ 1.57, p = 0.040). However, MR indicated negative certain subtypes of specific cancers, effect sizes for coronary heart lung (β − 4.759, 0.002), breast 2.684, 0.026), colorectal 4.581, 0.042), liver 19.264, 0.028), stroke prostate 0.299, 0.017), no evidence correlation. Results reverse revealed angina pectoris. An linked risk risk. has shown that expected incidence can reduce probability developing forms Further investigation required clinical correlations underlying processes these two illnesses.

Язык: Английский

Application of Mendelian randomization in thyroid diseases: a review

Frontiers in Endocrinology, Год журнала: 2024, Номер 15

Опубликована: Дек. 19, 2024

Thyroid diseases are increasingly prevalent, posing significant challenges to patients' quality of life and placing substantial financial burdens on families society. Despite these impacts, the underlying pathophysiology many thyroid conditions remains poorly understood, complicating efforts in treatment, management, prevention. Observational studies can identify associations between exposure variables disease; however, they often struggle account for confounding factors reverse causation. Understanding disease occurrence, epidemiological trends, clinical diagnosis, prevention, treatment relies heavily robust etiological research. Mendelian randomization, a method grounded genetics epidemiology, has been widely employed studying etiology diseases, offering solution some challenges. This paper categorizes into dysfunction cancer, reviewing related randomization studies. It further provides novel perspectives approaches investigating mechanisms designing intervention strategies.

Язык: Английский