Proceedings. IEEE International Conference on Bioinformatics and Biomedicine最新文献

Goal: Identifying population differences can serve as an insightful tool for diagnostic radiology. To do so, a reliable preprocessing framework and data representation are vital.

Methods: We build a machine learning model to visualize gender differences in the circle of Willis (CoW), an integral part of the brain's vasculature. We start with a dataset of 570 individuals and process them for analysis using 389 for the final analysis.

Results: We find statistical differences between male and female patients in one image plane and visualize where they are. We can see differences between the right and left-hand sides of the brain confirmed using Support Vector Machines (SVM).

Conclusion: This process can be applied to detect population variations in the vasculature automatically.

Significance: It can guide debugging and inferring complex machine learning algorithms such as SVM and deep learning models.

User-generated social media posts such as tweets can provide insights about the public's perception, cognitive, and behavioral responses to health-related issues. Pre-Exposure Prophylaxis (PrEP) is one of the most effective ways to reduce the risk of HIV infection. However, its utilization is low in the US, especially among populations disproportionately affected by HIV such as the age group of under 24 years old. It is therefore important to understand the barriers to the wider use of PrEP in the US using social media posts. In this study, we collected tweets from Twitter about PrEP in the past 4 years to identify such barriers by first identifying tweets about personal discussions, and then performing textual analysis using word analysis, UMLS semantic type analysis, and topic modeling. We found that the public often discussed advocacy, risks/benefits, access, pricing, insurance coverage, legislation, stigma, health education, and prevention of HIV. This result is consistent with the literature and can help identify strategies for promoting the use of PrEP, especially among young adults.

Accuracy of peptide identification in LC-MS analysis is crucial for information regarding the aspects of proteins that aid in biomarker discovery and the profiling of complex proteomes. The detection of peptide fragment ions in tandem mass spectrometry is still challenging given that current tools were not created or tested for the low-abundance, low-peak fragments of peptides found in MS2 data. Feature detection, a crucial pre-processing step in the LC-MS analysis pipeline that quantifies peptides by their mass-to-charge ratio, retention time, and intensity, is particularly challenging due to the overlapping nature of peptides and weak signals that are often indistinguishable from noises, thus creating a reliance on rigid mathematical structures and heuristics. In this study, we developed a deep-learning-based model with an innovative sliding window process that enables high-resolution processing of quantitative MS/MS data to conduct MS2 feature detection. Experimental results show that our model can produce more accurate values and identifications than existing feature detection tools, as well as a high rate of true positive features quantified. Therefore, we believe that our model illustrates the advantages of deep learning techniques applied towards computational proteomics.

Microbial community proteomics, also termed metaproteomics, investigates all proteins expressed by a microbiota. Tandem mass spectrometry (MS/MS) is the typical method for identifying proteins in metaproteomics, which involves searching the mass spectra against a protein sequence database. A major post-analysis step is controlling the false discovery rate (FDR), i.e., the ratio of false positives to the total number of annotations. The current popular target-decoy FDR estimation method treats all the peptides and proteins equally and overlooks that they could have varied probabilities of being identified. In this study, we report FineFDR, a framework for FDR assessment at fine-grained levels with taxonomy information considered. FineFDR groups the identified peptide-spectrum matches, peptides, and proteins from different taxonomic units and estimates the FDR in each group separately. Empirical experiments on the simulated and real-world data sets demonstrate that our FineFDR achieved higher precision and more peptide and protein identifications when compared to the state-of-the-art methods, such as Comet, Percolator, TIDD, and Tailor. FineFDR is freely available under the GNU GPL license at https://github.com/Biocomputing-Research-Group/FDR.

Electronic health records (EHR) have been widely applied to various tasks in the medical domain such as risk predictive modeling, which aims to predict further health conditions by analyzing patients' historical EHR. Existing work mainly focuses on modeling the sequential and temporal characteristics of EHR data with advanced deep learning techniques. However, the network architectures of these models are all manually designed based on experts' prior knowledge, which largely impedes non-experts from exploring this task. To address this issue, in this paper, we propose a novel automated risk prediction model named AutoMed to automatically search the optimal model architecture for modeling the complex EHR data and improving the performance of the risk prediction task. In particular, we follow the idea of neural architecture search to design a search space that contains three separate searchable modules. Two of them are used for analyzing sequential and temporal features of EHR data, respectively. The third is to automatically fuse both features together. Besides these three modules, AutoMed contains an embedding module and a prediction module. All the three searchable modules are jointly optimized in the search stage to derive the optimal model architecture. In such a way, the model design can be automatically achieved with few human interventions. Experimental results on three real-world datasets show that AutoMed outperforms state-of-the-art baselines in terms of PR-AUC, F1, and Cohen's Kappa. Moreover, the ablation study shows that AutoMed can obtain reasonable model architectures and offer useful insights to the future risk prediction model design.

Rapid accumulation of temporal Electronic Health Record (EHR) data and recent advances in deep learning have shown high potential in precisely and timely predicting patients' risks using AI. However, most existing risk prediction approaches ignore the complex asynchronous and irregular problems in real-world EHR data. This paper proposes a novel approach called Knowledge-guIded Time-aware LSTM (KIT-LSTM) for continuous mortality predictions using EHR. KIT-LSTM extends LSTM with two time-aware gates and a knowledge-aware gate to better model EHR and interprets results. Experiments on real-world data for patients with acute kidney injury with dialysis (AKI-D) demonstrate that KIT-LSTM performs better than the state-of-the-art methods for predicting patients' risk trajectories and model interpretation. KIT-LSTM can better support timely decision-making for clinicians.

The study of complex behavior of biological systems has become increasingly dependent on evolutionary network modeling. In particular, multi-omics networks capture interactions between biomolecules such as proteins and metabolites, providing a basis for predicting relationships between such biomolecules and various phenotypic traits of complex diseases. In this paper, we introduce an integrative framework that given a multi-omics network representing a cohort of subjects, learns expressive representations for network nodes, and combines the learned nodes representations with the biological profiles of individual subjects for enriched representation of the subjects. With extensive empirical evaluation using real-world multi-omics networks, we show that our proposed framework significantly outperforms existing and baseline methods in terms of subject representation accuracy, particularly when the multi-omics network representing the cohort is sparse and structured and therefore, more informative.

Successful prediction of clinical outcomes facilitates tailored diagnosis and treatment. The microbiome has been shown to be an important biomarker to predict host clinical outcomes. Further, the incorporation of microbial phylogeny, the evolutionary relationship among microbes, has been demonstrated to improve prediction accuracy. We propose a phylogeny-driven deep neural network (PhyNN) and develop an ensemble method, DeepEn-Phy, for host clinical outcome prediction. The method is designed to optimally extract features from phylogeny, thereby take full advantage of the information in phylogeny while harnessing the core principles of phylogeny (in contrast to taxonomy). We apply DeepEn-Phy to a real large microbiome data set to predict both categorical and continuous clinical outcomes. DeepEn-Phy demonstrates superior prediction performance to existing machine learning and deep learning approaches. Overall, DeepEn-Phy provides a new strategy for designing deep neural network architectures within the context of phylogeny-constrained microbiome data.

Although the U.S. FDA has only approved exactly one cannabidiol (CBD) drug product (specifically to treat seizures), CBD products are proliferating rapidly through different modes of usage including food products, cosmetics, vaping pods, and supplements (typically, oils). Despite the FDA clearly warning consumers about unproven health claims made by manufacturers selling CBD products over the counter, the CBD market share was nearly 3 billion USD in 2020 and is expected to top 55 billion USD in 2028. In this context, it is important to assess the presence of health claims being made on social media, especially claims that are part of marketing messages. To this end, we collected over two million English tweets discussing CBD themes. We created a hand-labeled dataset and built machine learned classifiers to identify marketing tweets from regular tweets that may be generated by consumers. The best classifier achieved 85% precision, 83% recall, and 84% F-score. Our analyses showed that pain, anxiety disorders, sleep disorders, and stress are the four main therapeutic claims made constituting 31.67%, 27.11%, 13.77%, and 10.37% of all medical claims made on Twitter, respectively. Also, more than 93% of advertised CBD products are edibles or oil/tinctures. Our effort is the first to demonstrate the feasibility of surveillance of marketing claims for CBD products. We believe this could pave way for more explorations into this indispensable task in the current landscape of social media driven health (mis)information and communication.

Alzheimer's disease (AD) is a progressive neurodegenerative brain disorder characterized by memory loss and cognitive decline. Early detection and accurate prognosis of AD is an important research topic, and numerous machine learning methods have been proposed to solve this problem. However, traditional machine learning models are facing challenges in effectively integrating longitudinal neuroimaging data and biologically meaningful structure and knowledge to build accurate and interpretable prognostic predictors. To bridge this gap, we propose an interpretable graph neural network (GNN) model for AD prognostic prediction based on longitudinal neuroimaging data while embracing the valuable knowledge of structural brain connectivity. In our empirical study, we demonstrate that 1) the proposed model outperforms several competing models (i.e., DNN, SVM) in terms of prognostic prediction accuracy, and 2) our model can capture neuroanatomical contribution to the prognostic predictor and yield biologically meaningful interpretation to facilitate better mechanistic understanding of the Alzheimer's disease. Source code is available at https://github.com/JaesikKim/temporal-GNN.