Machine Learning Approach to Detect Cardiac Arrhythmias in ECG Signals: A Survey

Cardiac arrhythmia is a condition when the heart rate is irregular either the beat is too slow or too fast. It occurs due to improper electrical impulses that coordinates the heart beats. Sudden cardiac death may occurs due to some dangerous arrhythmias conditions. Hence the main objective of the electrocardiogram (ECG) analysis is to detect the life-threatening arrhythmias accurately for appropriate treatment in order to save life. Since the last decades, several methods were reported for automatic ECG beat classifications. In this work, we present a systematic review of the current state-of-the-art methods used to detect cardiac arrhythmia using on ECG signals. It includes the signal decomposition, feature extraction and machine learning approaches used for automatic detection and decision making process. The articles covers the pre-processing, detection of QRS complex, feature extraction and classification of ECG beats. Based on the past studies, it is understood that the automated approach using computer-aided decision making process is highly required for real-time detection of cardiac arrhythmias. The advantages and limitations of different methods are discussed and also the future scopes is highlighted in the process of effective detection of cardiac arrhythmias. This study could be beneficial for researchers to analyze the existing state-of-art techniques used in detection of arrhythmia conditions.     

基于植保大数据的病虫害移动智能采集新设备

为提高农作物重大病虫害发生信息自动化、智能化采集能力,全面提升监测预警水平,笔者基于大数据、人工智能和深度学习技术,研发了一款农作物病虫害移动智能采集设备——智宝,主要实现了3个方面的功能:一是病虫害发生信息自动采集上报.通过该产品进行人工拍照,可实现对田间农作物重大病虫害发生图像、发生位置、发生数量、微环境因子等数据的实时采集和上报.二是自动识别计数.基于植保大数据与人工智能技术,通过构建病虫害自动识别系统,可实现重大病虫害精准识别与分析,只要拍摄照片,即可快速、精确地识别病虫害种类,并自动计数、上报到指定的测报系统.三是自动分析判别分级.针对拍摄采集上报的重大病虫害发生信息,系统可在自动识别和计数的基础上,进一步对病虫害发生严重程度进行智能判别分级,甚至根据相关预测模型,对病虫害的发生趋势进行辅助分析预测,提出预测建议.通过2016—2019年组织多地植保机构进行试验改进,该技术产品日趋成熟,有望在未来的农作物病虫害发生信息采集和预测预报工作中推广使用.     

Five candidate biomarkers associated with the diagnosis and prognosis of cervical cancer

Purpose: Cervical cancer (CC) is one of the most general gynecological malignancies and is associated with high morbidity and mortality. We aimed to select candidate genes related to the diagnosis and prognosis of CC.Methods: The mRNA expression profile datasets were downloaded. We also downloaded RNA-sequencing gene expression data and related clinical materials from TCGA, which included 307 CC samples and 3 normal samples. Differentially expressed genes (DEGs) were obtained by R software. GO function analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of DEGs were performed in the DAVID dataset. Using machine learning, the optimal diagnostic mRNA biomarkers for CC were identified. We used qRT-PCR and Human Protein Atlas (HPA) database to exhibit the differences in gene and protein levels of candidate genes.Results: A total of 313 DEGs were screened from the microarray expression profile datasets. DNA methyltransferase 1 (DNMT1), Chromatin Assembly Factor 1, subunit B (CHAF1B), Chromatin Assembly Factor 1, subunit A (CHAF1A), MCM2, CDKN2A were identified as optimal diagnostic mRNA biomarkers for CC. Additionally, the GEPIA database showed that the DNMT1, CHAF1B, CHAF1A, MCM2 and CDKN2A were associated with the poor survival of CC patients. HPA database and qRT-PCR confirmed that these genes were highly expressed in CC tissues.Conclusion: The present study identified five DEmRNAs, including DNMT1, CHAF1B, CHAF1A, MCM2 and Kinetochore-related protein 1 (KNTC1), as potential diagnostic and prognostic biomarkers of CC.     

Unwrapping the phase portrait features of adventitious crackle for auscultation and classification: a machine learning approach

Journal of Biological Physics - The paper delves into the plausibility of applying fractal, spectral, and nonlinear time series analyses for lung auscultation. The thirty-five sound signals of...     

基于机器学习的肠道菌群数据建模与分析研究综述

人体肠道菌群与人类的健康和疾病存在密切关系,对肠道菌群的宏基因组数据进行建模和分析,在疾病预测及诊断相关领域科学研究和社会应用方面均具有重要意义。本文从大数据分析和机器学习的角度,对人体肠道菌群数据的建模、分析和预测算法的原理、过程以及典型研究应用实例进行综述,以期推动肠道菌群分析相关研究发展以及探索结合机器学习算法进行肠道菌群分析的有效方式,同时也为开发基于肠道菌群数据的新型诊疗手段提供借鉴,推动我国精准医疗事业发展。     

亚致死剂量吡虫啉胁迫对意大利蜜蜂工蜂嗅觉学习行为的影响及其脑部转录组分析

【目的】分析吡虫啉处理对意大利蜜蜂Apis mellifera ligustica嗅觉学习行为及脑部基因转录的影响,为新烟碱类杀虫剂对蜜蜂的负面影响提供依据。【方法】实验室条件下一次性饲喂意大利蜜蜂成年工蜂含有4 ng吡虫啉的50%蔗糖溶液,以饲喂不含吡虫啉的50%蔗糖溶液为对照,通过伸吻反应(proboscis extension response, PER)行为实验测定其对意大利蜜蜂成年工蜂嗅觉学习行为的影响;收集上述测试的意大利蜜蜂工蜂脑部提取RNA,进行RNA-Seq测序和生物信息学分析,利用实时荧光定量PCR挑选6个差异表达基因(DEGs)检测其在脑部的表达量以验证RNA-Seq测序结果。【结果】一次性饲喂意大利蜜蜂成年工蜂含4 ng吡虫啉的50%蔗糖溶液后,意大利蜜蜂的嗅觉学习能力与对照组(饲喂50%蔗糖溶液)相比显著降低。RNA-Seq测序结果显示饲喂意大利蜜蜂含4 ng吡虫啉蔗糖溶液后,处理组与对照组之间共有123个DEGs［校正后的P值(padj)<0.05)］,包括82个下调DEGs和41个上调DEGs。GO聚类分析发现下调DEGs主要富集在S-腺苷甲硫氨酸依赖性甲基转移酶活性、酸性磷酸酶活性、氧化还原酶活性、蛋白质异二聚化活性等,上调DEGs主要富集在跨膜受体活性、分子传感器活性、神经学系统过程等功能条目。KEGG富集分析显示下调DEGs主要富集在核糖体和溶酶体等细胞器、碳代谢和色氨酸代谢等代谢通路及Toll和Imd信号通路,上调DEGs未富集在KEGG代谢通路。荧光定量PCR结果显示测试的6个DEGs相对表达量变化趋势与RNA-Seq测序结果FPKM(fragments per kilobase million)值变化趋势一致,证明RNA-Seq测序结果可靠。【结论】亚致死剂量吡虫啉胁迫意大利蜜蜂成年工蜂显著降低意大利蜜蜂的嗅觉学习能力,影响意大利蜜蜂脑部免疫解毒等相关基因的表达和酶活性及氧化还原等生物代谢过程;亚致死剂量吡虫啉短时胁迫会刺激意大利蜜蜂的嗅觉感觉过程及神经信号传导过程。     

基于F3Net显著性目标检测的蝴蝶图像前背景自动分割

【目的】具有复杂背景的蝴蝶图像前背景分割难度大。本研究旨在探索基于深度学习显著性目标检测的蝴蝶图像自动分割方法。【方法】应用DUTS-TR数据集训练F³Net显著性目标检测算法构建前背景预测模型,然后将模型用于具有复杂背景的蝴蝶图像数据集实现蝴蝶前背景自动分割。在此基础上,采用迁移学习方法,保持ResNet骨架不变,利用蝴蝶图像及其前景蒙板数据,使用交叉特征模块、级联反馈解码器和像素感知损失方法重新训练优化模型参数,得到更优的自动分割模型。同时,将其他5种基于深度学习显著性检测算法也用于自动分割,并比较了这些算法和F³Net算法的性能。【结果】所有算法均获得了很好的蝴蝶图像前背景分割效果,其中,F³Net是更优的算法,其7个指标S测度、E测度、F测度、平均绝对误差(MAE)、精度、召回率和平均IoU值分别为0.940, 0.945, 0.938, 0.024, 0.929,0.978和0.909。迁移学习则进一步提升了F³Net的上述指标值,分别为0.961, 0.964, 0.963, 0.013, 0.965, 0.967和0.938。【结论】研究结果证明结合迁移学习的F³Net算法是其中最优的分割方法。本研究提出的方法可用于野外调查中拍摄的昆虫图像的自动分割,并拓展了显著性目标检测方法的应用范围。     

Contrasting Gene Decay in Subterranean Vertebrates: Insights from Cavefishes and Fossorial Mammals

Evolution sometimes proceeds by loss, especially when structures and genes become dispensable after an environmental shift relaxes functional constraints. Subterranean vertebrates are outstanding models to analyze this process, and gene decay can serve as a readout. We sought to understand some general principles on the extent and tempo of the decay of genes involved in vision, circadian clock, and pigmentation in cavefishes. The analysis of the genomes of two Cuban species belonging to the genus Lucifuga provided evidence for the largest loss of eye-specific genes and nonvisual opsin genes reported so far in cavefishes. Comparisons with a recently evolved cave population of Astyanax mexicanus and three species belonging to the Chinese tetraploid genus Sinocyclocheilus revealed the combined effects of the level of eye regression, time, and genome ploidy on eye-specific gene pseudogenization. The limited extent of gene decay in all these cavefishes and the very small number of loss-of-function mutations per pseudogene suggest that their eye degeneration may not be very ancient, ranging from early to late Pleistocene. This is in sharp contrast with the identification of several vision genes carrying many loss-of-function mutations in ancient fossorial mammals, further suggesting that blind fishes cannot thrive more than a few million years in cave ecosystems.