The rapid growth of omics-based data has revolutionized biomedical research and precision medicine, allowing machine learning models to be developed for cutting-edge performance. However, despite the wealth of high-throughput data available, the performance of these models is hindered by the lack of sufficient training data, particularly in clinical research (in vivo experiments). As a result, translating this knowledge into clinical practice, such as predicting drug responses, remains a challenging task. Transfer learning is a promising tool that bridges the gap between data domains by transferring knowledge from the source to the target domain. Researchers have proposed transfer learning to predict clinical outcomes by leveraging pre-clinical data (mouse, zebrafish), highlighting its vast potential. In this work, we present a comprehensive literature review of deep transfer learning methods for health informatics and clinical decision-making, focusing on high-throughput molecular data. Previous reviews mostly covered image-based transfer learning works, while we present a more detailed analysis of transfer learning papers. Furthermore, we evaluated original studies based on different evaluation settings across cross-validations, data splits and model architectures. The result shows that those transfer learning methods have great potential; high-throughput sequencing data and state-of-the-art deep learning models lead to significant insights and conclusions. Additionally, we explored various datasets in transfer learning papers with statistics and visualization.

MOTIVATION: The early detection of cancer through accessible blood tests can foster early patient interventions. Although there are developments in cancer detection from cell-free DNA (cfDNA), its accuracy remains speculative. Given its central importance with broad impacts, we aspire to address the challenge. METHOD: A bagging Ensemble Meta Classifier (CancerEMC) is proposed for early cancer detection based on circulating protein biomarkers and mutations in cfDNA from blood. CancerEMC is generally designed for both binary cancer detection and multi-class cancer type localization. It can address the class imbalance problem in multi-analyte blood test data based on robust oversampling and adaptive synthesis techniques. RESULTS: Based on the clinical blood test data, we observe that the proposed CancerEMC has outperformed other algorithms and state-of-the-arts studies (including CancerSEEK) for cancer detection. The results reveal that our proposed method (i.e. CancerEMC) can achieve the best performance result for both binary cancer classification with 99.17% accuracy (AUC = 0.999) and localized multiple cancer detection with 74.12% accuracy (AUC = 0.938). Addressing the data imbalance issue with oversampling techniques, the accuracy can be increased to 91.50% (AUC = 0.992), where the state-of-the-art method can only be estimated at 69.64% (AUC = 0.921). Similar results can also be observed on independent and isolated testing data. AVAILABILITY: https://github.com/saifurcubd/Cancer-Detection. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Acute coronary syndrome (ACS) is responsible for the obstruction of coronary arteries, resulting in the loss of lives. The onset of ACS can be determined by looking at the various signs and symptoms of a patient. However, the accuracy of ACS determination is often put into question since there exist different types of uncertainties with the signs and symptoms. Belief rule-based expert systems (BRBESs) are widely used to capture uncertain knowledge and to accomplish the task of reasoning under uncertainty by employing belief rule base and evidential reasoning. This article presents the process of developing a BRBES to determine ACS predictability. The BRBES has been validated against the data of 250 patients suffering from chest pain. It is noticed that the outputs created from the BRBES are more dependable than that of the opinion of cardiologists as well as other two expert system tools, namely artificial neural networks and support vector machine. Hence, it can be argued that the BRBES is capable of playing an important role in decision making as well as in avoiding costly laboratory investigations. A procedure to train the system, allowing its enhancement of performance, is also presented.

A rapidly emerging trend in the IT landscape is the uptake of large-scale datacenters moving storage and data processing to providers located far away from the end-users or locally deployed servers. For these large-scale datacenters, power efficiency is a key metric, with the PUE (Power Usage Effectiveness) and DCiE (Data Centre infrastructure Efficiency) being important examples. This article proposes a belief rule based expert system to predict datacenter PUE under uncertainty. The system has been evaluated using real-world data from a data center in the UK. The results would help planning construction of new datacenters and the redesign of existing datacenters making them more power efficient leading to a more sustainable computing environment. In addition, an optimal learning model for the BRBES demonstrated which has been compared with ANN and Genetic Algorithm; and the results are promising.