当前共找到 1280 篇文献分享,本页显示第 321 - 340 篇。
321.
盼盼
(2024-04-30 22:50):
#paper doi:https://doi.org/10.1038/s41587-021-00830-w Robust decomposition of cell type mixtures in spatial transcriptomics
空间转录组学技术的局限在于,每个spot的测序数值可能来自于不同细胞的贡献,这样不利于细胞特异性和空间表达模式特异性的挖掘。本篇文章的作者开发了一个稳定性比较高的软件:RCTD,它利用从单细胞数据中细胞特异性谱图的表达水平,预测每个spot中细胞类型,并计算出每种细胞的权重。RCTD计算了小鼠Slide-seq跟visium数据集中准确的再现了已知的细胞类型和亚型细胞定位模式。不过这个方法结果的可靠程度依赖于注释好的单细胞数据集的质量,因此选择质量好的单细胞数据集,或者细胞注释准确度高的与空间数据匹配好的单细胞数据集是非常重要的。选择RCTD对空间数据spot的细胞类型的空间成分,揭示生物组织中细胞组织的新原理。
Abstract:
A limitation of spatial transcriptomics technologies is that individual measurements may contain contributions from multiple cells, hindering the discovery of cell-type-specific spatial patterns of localization and expression. Here, we develop …
>>>
A limitation of spatial transcriptomics technologies is that individual measurements may contain contributions from multiple cells, hindering the discovery of cell-type-specific spatial patterns of localization and expression. Here, we develop robust cell type decomposition (RCTD), a computational method that leverages cell type profiles learned from single-cell RNA-seq to decompose cell type mixtures while correcting for differences across sequencing technologies. We demonstrate the ability of RCTD to detect mixtures and identify cell types on simulated datasets. Furthermore, RCTD accurately reproduces known cell type and subtype localization patterns in Slide-seq and Visium datasets of the mouse brain. Finally, we show how RCTD's recovery of cell type localization enables the discovery of genes within a cell type whose expression depends on spatial environment. Spatial mapping of cell types with RCTD enables the spatial components of cellular identity to be defined, uncovering new principles of cellular organization in biological tissue. RCTD is publicly available as an open-source R package at https://github.com/dmcable/RCTD .
<<<
翻译
322.
尹志
(2024-04-30 22:48):
#paper doi:https://doi.org/10.48550/arXiv.2211.07697,NeurIPS 2022 Workshop on Symmetry and Geometry in Neural Representations, 2022. Do Neural Networks Trained with Topological Features Learn Different Internal Representations? 作者主要讨论了使用拓扑特征训练神经网络和使用常规数据直接进行神经网络训练在表征上的区别。结论很有意思,比较容易猜到的是,两者确实有区别,特别是在作者选择的metrics下,这也说明了拓扑机器学习的价值。但作者发现在一些情况下,也存在可以利用简单的表征来替代拓扑特征训练的模型。当然,在具体的数据场景下怎么样提取出合适的拓扑特征显著区别于使用raw data可以提取的特征,这仍是一个开放的主题。
arXiv,
2022.
DOI: 10.48550/arXiv.2211.07697
Abstract:
There is a growing body of work that leverages features extracted viatopological data analysis to train machine learning models. While this field,sometimes known as topological machine learning (TML), has seen …
>>>
There is a growing body of work that leverages features extracted viatopological data analysis to train machine learning models. While this field,sometimes known as topological machine learning (TML), has seen some notablesuccesses, an understanding of how the process of learning from topologicalfeatures differs from the process of learning from raw data is still limited.In this work, we begin to address one component of this larger issue by askingwhether a model trained with topological features learns internalrepresentations of data that are fundamentally different than those learned bya model trained with the original raw data. To quantify ``different'', weexploit two popular metrics that can be used to measure the similarity of thehidden representations of data within neural networks, neural stitching andcentered kernel alignment. From these we draw a range of conclusions about howtraining with topological features does and does not change the representationsthat a model learns. Perhaps unsurprisingly, we find that structurally, thehidden representations of models trained and evaluated on topological featuresdiffer substantially compared to those trained and evaluated on thecorresponding raw data. On the other hand, our experiments show that in somecases, these representations can be reconciled (at least to the degree requiredto solve the corresponding task) using a simple affine transformation. Weconjecture that this means that neural networks trained on raw data may extractsome limited topological features in the process of making predictions.
<<<
翻译
323.
孤舟蓑笠翁
(2024-04-30 22:44):
#paper doi:10.1101/2024.03.18.585576,bioRxiv,2024-03-19。Single-cell genomics and regulatory networks for 388 human brains。这个研究首次在人群规模对人脑前额叶区域进行了单细胞核转录组、染色质可及性测序,然后在细胞类型的精度对基因调控网络、细胞通讯网络等方面进行了生理和病理条件下的探究。研究结果可以在项目(brainSCOPE)的官网获取。官网:http://brainscope.psychencode.org。该研究用了388个人的脑。其中333个是该研究产生的,55个是外来的;健康个体有182个,其余有精神分裂症、双相障碍(抑郁狂躁型忧郁症)、自闭症或老年痴呆。388个个体有snRNA-seq数据。59个个体有snATAC-seq数据,其中40个的是snMultiome(对同一个细胞既测转录组又测ATAC)。质控后共280万个细胞核(注释到了28种细胞)。【研究角度及部分主要发现】1,对每种细胞找cis-eQTL和cis调控元件。2,构建细胞类型特异性的基因调控网络和细胞间通信网络,并展示这些网络在衰老和神经精神疾病中的变异。3,探究每种细胞的占比、基因表达、表观遗传和年龄、老年痴呆的关联。用基因表达量构建预测年龄的摸型。发现有6种细胞的转录组有很强的预测能力。4,在每种细胞里构建摸型,用遗传变异预测对细胞、组织的基因表达的影响。模拟基因序列的干绕对基因表达、表型(包括疾病倾向)等下游的影响。【研究的不足或未来研究方向】 1,RNA表达量不能代替蛋白表达量。这在某些脑区尤其突出。2,人去世后的脑组织和活人的脑组织有区别。3,研究更多脑区,以及发育、衰老中的脑区或者类器官。4,整合更多类型的数据,比如成像数据,用于提升预测表型的能力。【应用前景】1,为理解神经精神疾病的分子机制提供了新的视角,有助于发现新的治疗方法。2,通过整合模型(LNCTP),可以从基因型数据中预测个体的细胞类型特异性功能基因表达,为精准医疗提供工具。3,研究结果可用于优先考虑潜在的药物靶点,并模拟特定基因的表达变化,以预测其对疾病表型的潜在影响。4,该研究创建的brainSCOPE资源库可供其他研究者使用,以进一步探索大脑的分子结构和功能。总体而言,这项研究通过大规模的单细胞分析,为理解人类大脑的复杂性、疾病机制和潜在的治疗干预提供了宝贵的资源和新的洞见。
bioRxiv : the preprint server for biology,
2024-Mar-30.
DOI: 10.1101/2024.03.18.585576
PMID: 38562822
Abstract:
Single-cell genomics is a powerful tool for studying heterogeneous tissues such as the brain. Yet, little is understood about how genetic variants influence cell-level gene expression. Addressing this, we uniformly …
>>>
Single-cell genomics is a powerful tool for studying heterogeneous tissues such as the brain. Yet, little is understood about how genetic variants influence cell-level gene expression. Addressing this, we uniformly processed single-nuclei, multi-omics datasets into a resource comprising >2.8M nuclei from the prefrontal cortex across 388 individuals. For 28 cell types, we assessed population-level variation in expression and chromatin across gene families and drug targets. We identified >550K cell-type-specific regulatory elements and >1.4M single-cell expression-quantitative-trait loci, which we used to build cell-type regulatory and cell-to-cell communication networks. These networks manifest cellular changes in aging and neuropsychiatric disorders. We further constructed an integrative model accurately imputing single-cell expression and simulating perturbations; the model prioritized ~250 disease-risk genes and drug targets with associated cell types.
<<<
翻译
324.
哪有情可长
(2024-04-30 22:30):
#paper doi:https://doi.org/10.1038/s41588-024-01715-9,Nature genetics, 11 April 2024. A pan-genome of 69 Arabidopsis thaliana accessions reveals a conserved genome structure throughout the global species range. 通过对来自全球不同地理起源的 72 个拟南芥种质进行SNP分析,确定了四个主要遗传群体:“欧洲”、“非洲”、“马德拉”和“亚洲”,以及三个“混合”种质。使用了长读取(PacBio HiFi,平均深度为 45×,和牛津纳米孔,平均深度为 67×)和短读取测序,结合参考导向的拼接和手动修正,为每个 72 个种质生成了基因组组装。69 个种质经确认为纯合子品系,为了解释基因组大小变异的潜在基因组特征,选择了最完整的 46 个组装,并分析了组装与基因组大小估计的比值以及着丝粒重复长度与着丝粒大小估计的比值。这些种质的组装大小范围从 130 到 148 Mb。着丝粒重复序列平均长度为 14 Mb,与组装大小高度相关。通过从个体组装的初始 TE 注释生成的 pan-TE 库注释了转座元件。TE 空间大小在基因组之间非常相似,长末端重复序列和 Helitrons 占据了最大的 TE 分数。拟南芥基因组大小变异主要由着丝粒重复长度主导,而 TE 只是次要贡献者。单个染色体的大小独立于彼此地演化。后续作者利用共线性分析,发现拟南芥的染色体臂存在高度的同源性,而在着丝粒附近存在大型的重排现象。后续对泛基因组进行基因家族分析,发现一些核心、软核心、可选和私有家族基因。
Abstract:
Although originally primarily a system for functional biology, Arabidopsis thaliana has, owing to its broad geographical distribution and adaptation to diverse environments, developed into a powerful model in population genomics. …
>>>
Although originally primarily a system for functional biology, Arabidopsis thaliana has, owing to its broad geographical distribution and adaptation to diverse environments, developed into a powerful model in population genomics. Here we present chromosome-level genome assemblies of 69 accessions from a global species range. We found that genomic colinearity is very conserved, even among geographically and genetically distant accessions. Along chromosome arms, megabase-scale rearrangements are rare and typically present only in a single accession. This indicates that the karyotype is quasi-fixed and that rearrangements in chromosome arms are counter-selected. Centromeric regions display higher structural dynamics, and divergences in core centromeres account for most of the genome size variations. Pan-genome analyses uncovered 32,986 distinct gene families, 60% being present in all accessions and 40% appearing to be dispensable, including 18% private to a single accession, indicating unexplored genic diversity. These 69 new Arabidopsis thaliana genome assemblies will empower future genetic research.
<<<
翻译
325.
朵朵
(2024-04-30 22:13):
#paper 齐腾飞,高良敏.中国援非医疗队:在新自由主义卫生泥潭中摸索前行[J].文化纵横,2024(02):132-141.
论文认为,与社会主义时代南方国家之间的人道主义援助不同,当前中国援外医疗队所面对的非洲医疗生态,已被数十年来的全球新自由主义潮流所形塑,这一方面保障了医生的执业自由、私立医院和医疗旅游的发展,另一方面却削弱了普通民众获取基本医疗需求的可能性。这种医疗生态折射的价值观与中国援非医疗队的定位相悖,导致援非医疗队员时常陷入窘境。
这与我在非洲调研时的观感相符。因为非洲医疗体制照搬西方,一些中国医生到了非洲却没有处方权,没有办法开门诊。等到办完所有手续,任期也差不多该结束了。医疗如此,环保政策亦是如此。是照搬新自由主义思想,还是根据国情选择自主道路,决策者需要非常清醒。
326.
muton
(2024-04-30 22:07):
#paper doi: 10.1126/science.adk8261. Epub 2024 Mar 28. Selection of experience for memory by hippocampal sharp wave ripples 记忆对人类来说是非常重要的,但是并不是所有经历过的事件都能够被记住,那么大脑是如何筛选出值得记住的瞬间的呢? 纽约大学的György Buzsáki教授发现一种神经振荡的信号叫做SWR,也就是尖波涟漪波,他们认为清醒时期尖波涟漪波的发放会帮助记忆“打标”,帮助我们从事件中选择出有用的部分,并且在本文中的设想是清醒时起有尖波涟漪波发放附近的事件在睡眠时期会同样发放SWR并以replay的形式进行记忆巩固,帮助增强了记忆。本文运用了序列非负矩阵分解(seqNMF) 与统一流形的估计投影技术(UMAP)帮助解码了小鼠在迷宫路径探索下的行为和时间信息,使这一假设得到了数据的支持。
Abstract:
Experiences need to be tagged during learning for further consolidation. However, neurophysiological mechanisms that select experiences for lasting memory are not known. By combining large-scale neural recordings in mice with …
>>>
Experiences need to be tagged during learning for further consolidation. However, neurophysiological mechanisms that select experiences for lasting memory are not known. By combining large-scale neural recordings in mice with dimensionality reduction techniques, we observed that successive maze traversals were tracked by continuously drifting populations of neurons, providing neuronal signatures of both places visited and events encountered. When the brain state changed during reward consumption, sharp wave ripples (SPW-Rs) occurred on some trials, and their specific spike content decoded the trial blocks that surrounded them. During postexperience sleep, SPW-Rs continued to replay those trial blocks that were reactivated most frequently during waking SPW-Rs. Replay content of awake SPW-Rs may thus provide a neurophysiological tagging mechanism to select aspects of experience that are preserved and consolidated for future use.
<<<
翻译
327.
符毓
(2024-04-30 21:53):
#paper doi:10.1002/eem2.12734, Energy&Environmental Materials, 2024, Confluence of ZnO and PTFE Binder for Enhancing Performance of Thin-Film Lithium-Ion Batteries。开发具有高比容量和循环稳定性的负极材料对于改进薄膜锂离子电池至关重要。 薄膜氧化锌(ZnO)由于其高比容量而具有前景,但它在循环过程中会受到体积变化和结构应力的影响,导致电池性能较差。本文用磁控溅射方法聚四氟乙烯(PTFE)与ZnO结合在一起,确保了薄膜复合电极的牢固结合。PTFE有效降低了活性材料上的应力并减轻了Li+离子嵌入和脱嵌过程中的体积变化影响。ZnO/PTFE薄膜电极从第1次到第100次循环的容量保持率高达82%,超过了裸ZnO薄膜的50%
Abstract:
Developing anode materials with high specific capacity and cycling stability is vital for improving thin‐film lithium‐ion batteries. Thin‐film zinc oxide (ZnO) holds promise due to its high specific capacity, but …
>>>
Developing anode materials with high specific capacity and cycling stability is vital for improving thin‐film lithium‐ion batteries. Thin‐film zinc oxide (ZnO) holds promise due to its high specific capacity, but it suffers from volume changes and structural stress during cycling, leading to poor battery performance. In this research, we ingeniously combined polytetrafluoroethylene (PTFE) with ZnO using a radio frequency (RF) magnetron co‐sputtering method, ensuring a strong bond in the thin‐film composite electrode. PTFE effectively reduced stress on the active material and mitigated volume change effects during Li+ ion intercalation and deintercalation. The composite thin films are thoroughly characterized using advanced techniques such as X‐ray diffraction, scanning electron microscopy, and X‐ray photoelectron spectroscopy for investigating correlations between material properties and electrochemical behaviors. Notably, the ZnO/PTFE thin‐film electrode demonstrated an impressive specific capacity of 1305 mAh g−1 (=7116 mAh cm−3) at a 0.5C rate and a remarkable capacity retention of 82% from the 1st to the 100th cycle, surpassing the bare ZnO thin film (50%). This study provides valuable insights into using binders to stabilize active materials in thin‐film batteries, enhancing battery performance.
<<<
翻译
328.
庞庞
(2024-04-30 21:44):
#paper doi:10.1038/s41591-023-02296-6 Heterogeneous aging across multiple organ systems and prediction of chronic disease and mortality
可能我们大家比较熟悉人脑的年龄,即使用机器学习模型,基于脑指标(如功能连接、灰质体积等)预测人的年龄,预测值如果比真实年龄高,说明这个人比同龄人的脑子更加老化,反之更加年轻。预测值与真实值的差值可以衡量一个人脑的老化程度。而在这里,作者则使用UKbiobank数据集,进一步收集了除脑子以外身体的数据,构建了各个器官系统的年龄,进一步探究了器官系统的年龄是如何互相影响的、以及是如何影响脑龄的,即构建了一个Multi-organ的网络。同时他们也探究了哪些生活因素与器官老化有关,还有各种慢性病的器官年龄是怎样的异常模式。这篇文章给我们提供了一个研究个体老化的新视角,很有创新性。
Abstract:
Biological aging of human organ systems reflects the interplay of age, chronic disease, lifestyle and genetic risk. Using longitudinal brain imaging and physiological phenotypes from the UK Biobank, we establish …
>>>
Biological aging of human organ systems reflects the interplay of age, chronic disease, lifestyle and genetic risk. Using longitudinal brain imaging and physiological phenotypes from the UK Biobank, we establish normative models of biological age for three brain and seven body systems. Here we find that an organ's biological age selectively influences the aging of other organ systems, revealing a multiorgan aging network. We report organ age profiles for 16 chronic diseases, where advanced biological aging extends from the organ of primary disease to multiple systems. Advanced body age associates with several lifestyle and environmental factors, leukocyte telomere lengths and mortality risk, and predicts survival time (area under the curve of 0.77) and premature death (area under the curve of 0.86). Our work reveals the multisystem nature of human aging in health and chronic disease. It may enable early identification of individuals at increased risk of aging-related morbidity and inform new strategies to potentially limit organ-specific aging in such individuals.
<<<
翻译
329.
钟鸣
(2024-04-30 17:57):
#paper doi:10.1177/0141076816688346 Prevention and treatment of diabetic foot ulcers
糖尿病足是糖尿病最常见且危害很大的并发症之一,表现为足底溃疡坏死,严重者需要截肢。本文综述了糖尿病足的预防和治疗。病因方面,大致可理解为异常的糖代谢产物诱发神经细胞氧化应激,这些受损的神经元引起了异常的足底压力分配、负重、步态、痛觉敏感性、汗腺功能、脚型等。干预的主要措施是预防,包括医护人员对足部护理的投入程度,以及患者自身被教育的程度。治疗方面,控制血糖是根本,以及及时有效的清创术(清除浅表及深部的坏死及过度角化组织)、药物治疗、各种各样促进伤口愈合的敷料、负压抽吸伤口组织液以改善灌注和促进肉芽形成、用蛆去除坏死组织、使用生长因子促进组织愈合等。
IF:8.800Q1
Journal of the Royal Society of Medicine,
2017-Mar.
DOI: 10.1177/0141076816688346
PMID: 28116957
Abstract:
The rising prevalence of diabetes estimated at 3.6 million people in the UK represents a major public health and socioeconomic burden to our National Health Service. Diabetes and its associated …
>>>
The rising prevalence of diabetes estimated at 3.6 million people in the UK represents a major public health and socioeconomic burden to our National Health Service. Diabetes and its associated complications are of a growing concern. Diabetes-related foot complications have been identified as the single most common cause of morbidity among diabetic patients. The complicating factor of underlying peripheral vascular disease renders the majority of diabetic foot ulcers asymptomatic until latter evidence of non-healing ulcers become evident. Therefore, preventative strategies including annual diabetic foot screening and diabetic foot care interventions facilitated through a multidisciplinary team have been implemented to enable early identification of diabetic patients at high risk of diabetic foot complications. The National Diabetes Foot Care Audit reported significant variability and deficiencies of care throughout England and Wales, with emphasis on change in the structure of healthcare provision and commissioning, improvement of patient education and availability of healthcare access, and emphasis on preventative strategies to reduce morbidities and mortality of this debilitating disease. This review article aims to summarise major risk factors contributing to the development of diabetic foot ulcers. It also considers the key evidence-based strategies towards preventing diabetic foot ulcer. We discuss tools used in risk stratification and classifications of foot ulcer.
<<<
翻译
330.
小年
(2024-04-30 13:46):
#paper genome-wide spectrum of tandem repeat expansions in 338,963 humans. Cell. 2024 Apr 25;187(9):2336-2341.e5. doi: 10.1016/j.cell.2024.03.004. Epub 2024 Apr 5. PMID: 38582080.
在本篇文章中,作者建立了一个名为TR-gnomAD的串联重复(TR)扩增的生物库规模参考数据库,该数据库涵盖了338,963个全基因组测序样本,其中包括39.5%的非欧洲样本。本研究使用了ExpansionHunter和GangSTR两种基因分型工具来增加TR基因分型的覆盖率。TR-gnomAD提供了关于TR单元数量频率差异的洞见,这些差异揭示了特定祖源与疾病风险之间的联系。例如,作者发现特定祖源中TR单元的扩增频率与某些遗传病的流行程度相关联,比如非洲后裔中肌阵挛型营养不良的DMPK基因中的CAG重复扩增较欧洲后裔中的较少。
阅读思考:TR-gnomAD的建立极大地丰富了我们对人类基因组中TR区域多样性的理解,并提供了一种全新的角度来观察遗传疾病与特定人种间的关联。这项研究强调了TR扩增在特定祖源中可能具有的疾病相关性,特别是在ALS、亨廷顿病等超过50种与TR扩增相关的致命性疾病中。此外,该数据库也对临床诊断提供了重要的参考,尤其是在使用TR 单元数差异来预测疾病风险方面。然而,该研究的一个限制是其主要依赖于短读测序数据,可能导致对大的TR扩增区域的等位基因长度估计不足。未来研究需结合长读测序技术,以提供更准确的TR扩增数据,从而更好地服务于遗传疾病的风险评估和诊断。
Abstract:
The Genome Aggregation Database (gnomAD), widely recognized as the gold-standard reference map of human genetic variation, has largely overlooked tandem repeat (TR) expansions, despite the fact that TRs constitute ∼6% …
>>>
The Genome Aggregation Database (gnomAD), widely recognized as the gold-standard reference map of human genetic variation, has largely overlooked tandem repeat (TR) expansions, despite the fact that TRs constitute ∼6% of our genome and are linked to over 50 human diseases. Here, we introduce the TR-gnomAD (https://wlcb.oit.uci.edu/TRgnomAD), a biobank-scale reference of 0.86 million TRs derived from 338,963 whole-genome sequencing (WGS) samples of diverse ancestries (39.5% non-European samples). TR-gnomAD offers critical insights into ancestry-specific disease prevalence using disparities in TR unit number frequencies among ancestries. Moreover, TR-gnomAD is able to differentiate between common, presumably benign TR expansions, which are prevalent in TR-gnomAD, from those potentially pathogenic TR expansions, which are found more frequently in disease groups than within TR-gnomAD. Together, TR-gnomAD is an invaluable resource for researchers and physicians to interpret TR expansions in individuals with genetic diseases.
<<<
翻译
331.
徐炳祥
(2024-04-30 13:16):
#paper doi: 10.1186/s13059-020-02167-0 Genome Biology, 2020, Mustache: multi-scale detection of chromatin loops from Hi-C and Micro-C maps using scale-space representation。染色质环是染色质空间构象的重要组成部分,也是启动子-增强子相互作用的重要物理背景。基于Hi-C数据的染色质环检测是当前三维基因组学的重要命题。本文立足于计算机视觉中的尺度稳定斑点检测技术开发了一种高灵敏度,高稳定的基于染色质相互作用图谱的染色质环检测算法。该算法是局部最大值搜索这一思路的最新作品,能在保证染色质换检测准确度的前体下大幅度提高其灵敏度。其综合性能为此类算法中最优者。
Abstract:
We present MUSTACHE, a new method for multi-scale detection of chromatin loops from Hi-C and Micro-C contact maps. MUSTACHE employs scale-space theory, a technical advance in computer vision, to detect …
>>>
We present MUSTACHE, a new method for multi-scale detection of chromatin loops from Hi-C and Micro-C contact maps. MUSTACHE employs scale-space theory, a technical advance in computer vision, to detect blob-shaped objects in contact maps. MUSTACHE is scalable to kilobase-resolution maps and reports loops that are highly consistent between replicates and between Hi-C and Micro-C datasets. Compared to other loop callers, such as HiCCUPS and SIP, MUSTACHE recovers a higher number of published ChIA-PET and HiChIP loops as well as loops linking promoters to regulatory elements. Overall, MUSTACHE enables an efficient and comprehensive analysis of chromatin loops. Available at: https://github.com/ay-lab/mustache .
<<<
翻译
332.
前进
(2024-04-30 11:44):
#paper Han D, Pan X, Han Y, et al. Flatten transformer: Vision transformer using focused linear attention[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. 2023: 5961-5971. 自注意力(self-attention)在计算机视觉任务中应用时面临的主要挑战是其二次计算复杂度,这使得处理视觉任务变得非常昂贵。作为Softmax注意力的一种替代方案,线性注意力通过精心设计的映射函数来近似Softmax操作,从而将计算复杂度从二次降低到线性。尽管线性注意力在理论上更加高效,但现有的线性注意力方法要么性能显著下降,要么需要额外的计算开销,这限制了它们的实际应用。为了克服这些限制,论文提出了FLA模块,它通过两个主要的改进来提高效率和表达能力:焦点能力:1 通过一个简单的映射函数,增强了自注意力对最信息特征的聚焦能力。特征多样性:引入了一个高效的秩恢复模块,通过深度卷积(DWC)来恢复注意力矩阵的秩,增加了特征的多样性。通过在多个先进的视觉Transformer模型上的广泛实验,FLA模块在多个基准测试中均显示出了一致的性能提升。
arXiv,
2023.
Abstract:
No abstract available.
333.
半面阳光
(2024-04-30 11:29):
#paper DOI: https://doi.org/10.1145/1273445.1273458 , ACM SIGCOMM Computer Communication Review, 2007, How to read a paper. 这是一篇讲如何如paper的paper。每个研究工作者都会花大量时间读paper,但读paper这项技能却很少被教授,从而导致因为读papar不得法而浪费大量时间。为了解决这个问题,作者在这篇papar中介绍了一种“三遍读papar”(three-pass method)法。此外,作者还简要介绍了进行文献调研的three-pass法,以及自己使用这些方法的个人经验和拓展资料。下面是对作者的“三遍读paper”方法的简要总结和整理。
【第一遍】快速扫描、通览全篇。大概花费5-10分钟。这一遍可以分为5个步骤。
(1)认真读title、abstract和introduction;
(2)读每个部分的标题和副标题,但是忽略其他内容;
(3)读conclusions;
(4)浏览references,在心里勾选出那些已经读过的文章。
读完这一遍,你应该能够回答“5C”问题。
C1:category,这是一篇什么类型的papar;
C2:context,这篇paper与其他那些papers有关,使用了哪些理论基础;
C3:Correctness,文章的假设是否成立;
C4: Contributions,这篇文章的主要贡献是什么;
C5:Clarity, 文章写得如何。
这一遍的阅读还可以给你写论文一些提醒,如果审稿人读一遍还没看到要点,这篇文章十有八九会被拒,如果读者读了5分钟还没看到重点,那多半不会再读这篇文章。
【第二遍】更加仔细,但不陷入细节(如,具体证据等),这一遍的目的是记下要点,在空白处记录一些要点评论。这一遍需要重点做两件事。
(1)认真阅读文章中的数据、图表和其他插图。特别需要注意图表中的坐标、误差、结论的统计结果是否可靠等。这些信息通常能快速区分好文章和差文章。
(2)标记相关的未读参考文献,以便进一步阅读。
这一遍最多花费1小时。
读完这一遍,你应该能够:把握文章的主要内容,可以向他人概括文章的主要内容并提供佐证。这一遍阅读的详略程度适合用于阅读你感兴趣但是不太熟悉的领域。
有时候,读完这一遍你可能并未理解文章的内容,这可能跟很多方面的原因有关。这时候,你可以把这个文章放在一边,期待自己不读不理解它也不影响职业生涯的成功;你也可以在了解一些背景知识后再读;也可以继续读下一遍。
【第三遍】对文章进行虚拟重现,也就是跟作者做出同样的假设,然后重新构建这项工作。把自己的重构与这篇文章进行比较,你可以轻松地发现这篇文章的创新之处,同时能够发现其隐藏的缺陷和假设。
这一遍要特别关注细节。你需要识别并挑战每一个观点背后的假设,你需要思考换做自己会如何表述某个观点。
这种模拟与实际的比较,可以对论文中的证明和表达技巧提供敏锐的洞察。在这个过程中,你还应记录下自己对未来工作的想法。
这一遍,初学者可能花费4-5小时,有经验的读者可能花费1小时。
读完这一遍,你应该能够:凭借记忆重构文章的所有内容,识别文章的优缺点。特别是能够指出隐藏假设、缺少的相关引用、实验或者分析方法可能存在的问题。
Abstract:
Researchers spend a great deal of time reading research papers. However, this skill is rarely taught, leading to much wasted effort. This article outlines a practical and efficient three-pass method …
>>>
Researchers spend a great deal of time reading research papers. However, this skill is rarely taught, leading to much wasted effort. This article outlines a practical and efficient three-pass method for reading research papers. I also describe how to use this method to do a literature survey.
<<<
翻译
334.
白鸟
(2024-04-30 09:13):
#paper doi:10.1016/j.phrs.2023.106800,Single-cell RNA-sequencing data reveals the genetic source of extracellular vesicles in esophageal squamous cell carcinoma, 2023.3. 分析恶性和非恶性食管组织中EV的遗传起源,揭示ESCC中细胞间相互作用;
文章策略:单细胞测序+外泌体RNA测序;
(1) 实验样本:6名ESCC患者同时进行单细胞测序,和提取不同细胞上清液,检测囊泡mRNA检测;
(2) 分析恶性和非恶性食管组织中EV的遗传起源,揭示ESCC中细胞间相互作用;
(3) 外泌体RNA测序-->揭示EV表达差异:恶性和非恶性食管组织中EV的基因表达存在差异;
(4) 单细胞测序-->精确识别EV的细胞起源:恶性组织主要是上皮细胞分泌EV,非恶性组织主要是内皮细胞和成纤维细胞分泌EV;
Abstract:
Esophageal squamous cell carcinoma (ESCC) is invasive cancer and the complex mechanisms underlying carcinogenesis remain unclear. Extracellular vesicles (EVs), secreted by most cell types, serve as a critical factor in …
>>>
Esophageal squamous cell carcinoma (ESCC) is invasive cancer and the complex mechanisms underlying carcinogenesis remain unclear. Extracellular vesicles (EVs), secreted by most cell types, serve as a critical factor in tumorigenesis via intercellular communications. Our study aims to investigate the cellular origin of EVs in ESCC, and unveil the unknown molecular and cellular mechanisms underlying cell-cell communications. Six ESCC patients were enrolled and single-cell RNA sequencing (scRNA-seq) analyses were conducted to screen different cell subpopulations. The genetic origin of EVs was tracked using the supernatant from different cellular extracts. Nanoparticle tracking analysis (NTA), western blot analysis, and transmission electron microscopy (TEM) were performed for validation. Using scRNA-seq analysis, eleven cell subpopulations were identified in ESCC. Differences in gene expression in EVs between malignant and non-malignant esophageal tissues were found. Our findings demonstrated that epithelial cells releasing EVs were the most prevalent in malignant tissues, while endothelial cells and fibroblasts releasing EVs were predominant in non-malignant tissues. Furthermore, the high levels of gene expression in EVs released from these cells were correlated significantly with a worse prognosis. Our findings revealed the genetic origin of EVs in malignant and non-malignant esophageal tissues and provided a comprehensive overview of the associated cell-cell interactions in ESCC.
<<<
翻译
335.
张浩彬
(2024-04-29 20:35):
#paper doi:
https://doi.org/10.48550/arXiv.2211.14730
A Time Series is Worth 64 Words: Long-term Forecasting with Transformers
ICLR2023的文章,提出了PatchTST。受vision Transformer的启发,把patch技术引入到时序问题。并且回应了早期另一篇认为Transformer用在时间序列其实并不比传统线性模型好的文章(Are transformers effective for time series forecasting?(2022)),重新取得了sota。然而23年底,又有新方法出现了,讨论了其实关键不是transformer,而是patch技术
arXiv,
2022.
DOI: 10.48550/arXiv.2211.14730
Abstract:
We propose an efficient design of Transformer-based models for multivariatetime series forecasting and self-supervised representation learning. It isbased on two key components: (i) segmentation of time series intosubseries-level patches which …
>>>
We propose an efficient design of Transformer-based models for multivariatetime series forecasting and self-supervised representation learning. It isbased on two key components: (i) segmentation of time series intosubseries-level patches which are served as input tokens to Transformer; (ii)channel-independence where each channel contains a single univariate timeseries that shares the same embedding and Transformer weights across all theseries. Patching design naturally has three-fold benefit: local semanticinformation is retained in the embedding; computation and memory usage of theattention maps are quadratically reduced given the same look-back window; andthe model can attend longer history. Our channel-independent patch time seriesTransformer (PatchTST) can improve the long-term forecasting accuracysignificantly when compared with that of SOTA Transformer-based models. We alsoapply our model to self-supervised pre-training tasks and attain excellentfine-tuning performance, which outperforms supervised training on largedatasets. Transferring of masked pre-trained representation on one dataset toothers also produces SOTA forecasting accuracy. Code is available at:https://github.com/yuqinie98/PatchTST.
<<<
翻译
336.
李翛然
(2024-04-28 18:09):
#paper doi:10. 1186/s42825-019-0012-x Nature Communication. Quantitative and structural analysis of isotopically labelled natural crosslinks in type I skin collagen using LC-HRMS and SANS 本文介绍了对使用LC-HRMS和SANS对标记同位素的天然交联物在I型皮肤胶原蛋白中进行定量和结构分析的研究。研究重点放在皮肤中的两种主要交联物HLNL和HHMD上,它们被同位素标记并进行分析,以了解它们的结构变化以及与硫酸铬的相互作用。研究强调了开发一种良性交联方法的重要性,以保留胶原蛋白的固有物理特性,特别是在皮革制造行业。主要发现包括确认HLNL和HHMD中各有一个亚胺基,使它们容易在低pH值下降解,并由于极端pH值变化和铬鞣制造导致胶原蛋白的结构变化。本研究使用的分析方法也可应用于研究其他胶原组织中的人工交联,用于生物医学应用。 这个算是人类第一篇弄清楚了胶原蛋白到底有哪些交联键~~所以化学交联的方法基本没戏,还是生物方法吧。~
Abstract:
Abstract Collagen structure in biological tissues imparts its intrinsic physical properties by the formation of several covalent crosslinks. For the first time, two major crosslinks in the skin dihydroxylysinonorleucine (HLNL) …
>>>
Abstract Collagen structure in biological tissues imparts its intrinsic physical properties by the formation of several covalent crosslinks. For the first time, two major crosslinks in the skin dihydroxylysinonorleucine (HLNL) and histidinohydroxymerodesmosine (HHMD), were isotopically labelled and then analysed by liquid-chromatography high-resolution accurate-mass mass spectrometry (LC-HRMS) and small-angle neutron scattering (SANS). The isotopic labelling followed by LC-HRMS confirmed the presence of one imino group in both HLNL and HHMD, making them more susceptible to degrade at low pH. The structural changes in collagen due to extreme changes in the pH and chrome tanning were highlighted by the SANS contrast variation between isotopic labelled and unlabelled crosslinks. This provided a better understanding of the interaction of natural crosslinks with the chromium sulphate in collagen suggesting that the development of a benign crosslinking method can help retain the intrinsic physical properties of the leather. This analytical method can also be applied to study artificial crosslinking in other collagenous tissues for biomedical applications. Graphical abstract
<<<
翻译
337.
颜林林
(2024-04-28 14:25):
#paper doi:10.1101/2022.11.29.518309, bioRxiv, 2024, NanoTrans: an integrated computational framework for comprehensive transcriptome analysis with Nanopore direct-RNA sequencing. 这篇预发表文章,开发了一套分析流程NanoTrans,用于Nanopore直接RNA测序(DRS)数据,进行全面的转录组分析,包括各基因及其转录本的聚类、定量、poly-A尾巴长度profiling、RNA修饰分析、融合基因检测等。文章本身在技术上并没有特别的创新,但将各方面的分析步骤,比较全面地整合到一起,提供一站式的功能封装,并以单HTML形式输出结果报告,这对于使用者还是很友好且很有用的。同时,文章在多种真实数据集(包括酵母、拟南芥、人胚胎肾和癌细胞系)上进行了测试,以证明其适用于不同的生物学应用场景。我个人觉得,这种流程开发的工作,其实很难发表得比较好(当经常地,我们又不得不花大量时间来做),想要进一步提升价值,需要更深入地在某些特定场景下进行改进和优化,而不是一味求全,但相应地,针对特定场景的数据所做的优化,会进一步限制流程软件的适用范围,这种时候如果结果不出彩(比如没有一些新奇发现),最终价值也同样会非常受限。
bioRxiv,
2022.
DOI: 10.1101/2022.11.29.518309
Abstract:
Nanopore direct RNA sequencing (DRS) provides the direct access to native RNA strands with full-length information, shedding light on rich qualitative and quantitative properties of gene expression profiles. Here with …
>>>
Nanopore direct RNA sequencing (DRS) provides the direct access to native RNA strands with full-length information, shedding light on rich qualitative and quantitative properties of gene expression profiles. Here with NanoTrans, we present an integrated computational framework that comprehensively covers all major DRS-based application scopes, including isoform clustering and quantification, poly(A) tail length estimation, RNA modification profiling, and fusion gene detection. In addition to its merit in providing such a streamlined one-stop solution, NanoTrans also shines in its workflow-orientated modular design, batch processing capability, all-in-one tabular and graphic report output, as well as automatic installation and configuration supports. Finally, by applying NanoTrans to real DRS datasets of yeast, Arabidopsis, as well as human embryonic kidney and cancer cell lines, we further demonstrated its utility, effectiveness, and efficacy across a wide range of DRS-based application settings.
<<<
翻译
338.
惊鸿
(2024-04-24 18:51):
#paper DNA origami mediated electrically connected metal—semiconductor junctions Pub Date : 2020-02-19
DOI : 10.1007/s12274-020-2672-5 无机纳米结构的基于DNA的纳米加工在电子,催化和等离激元学中具有潜在的应用。先前的DNA金属化已经产生了导电的DNA组装的纳米结构。然而,半导体的使用以及DNA纳米结构上良好连接的纳米级金属-半导体结的开发仍处于早期阶段。本文中,我们报告了通过金和碲纳米棒的位置特异性结合在单个DNA折纸上首次制造多个电连接的金属-半导体结。纳米棒附着到DNA折纸上的方法是通过DNA杂交获得Au,通过静电相互作用获得Te。化学镀金用于通过填充Au和Te纳米棒之间的间隙来创建纳米级金属-半导体界面。两点电特性表明,Au-Te-Au结已电连接,电流-电压特性与肖特基结一致。基于DNA的金属-半导体结的纳米加工为纳米电子学打开了潜在的机遇,证明了这种自下而上方法的强大功能。
Abstract:
No abstract available.
339.
DeDe宝
(2024-04-05 01:19):
#paper, Prior probability cues bias sensory encoding with increasing task exposure, doi:https://elifesciences.org/articles/91135.先验影响被试的反应时间和反应结果,之前的计算模型通常将先验的影响归因于选择方案时的调整(即先验的影响发生在决策阶段)。然而,最近的研究表明,先验也可能直接影响感官证据(似然)。在本研究中,研究者用脑电图记录了神经活动,同时被试在多个测试过程中使用有效先验、无效先验或中性先验执行对比辨别任务。研究者通过对比依赖的稳态视觉诱发电位(SSVEP)测量了感觉证据编码,而标准调整的读数则由运动皮层上的效应器选择性 mu-β 带活动提供。研究着发现先验对运动和SSVEP 均产生了显着的调节。研究表明,除了对决策过程进行战略调整之外,先验概率信息还会引起编码的偏差。
Abstract:
When observers have prior knowledge about the likely outcome of their perceptual decisions, they exhibit robust behavioural biases in reaction time and choice accuracy. Computational modelling typically attributes these effects …
>>>
When observers have prior knowledge about the likely outcome of their perceptual decisions, they exhibit robust behavioural biases in reaction time and choice accuracy. Computational modelling typically attributes these effects to strategic adjustments in the criterion amount of evidence required to commit to a choice alternative - usually implemented by a starting point shift - but recent work suggests that expectations may also fundamentally bias the encoding of the sensory evidence itself. Here, we recorded neural activity with EEG while participants performed a contrast discrimination task with valid, invalid, or neutral probabilistic cues across multiple testing sessions. We measured sensory evidence encoding via contrast-dependent steady-state visual-evoked potentials (SSVEP), while a read-out of criterion adjustments was provided by effector-selective mu-beta band activity over motor cortex. In keeping with prior modelling and neural recording studies, cues evoked substantial biases in motor preparation consistent with criterion adjustments, but we additionally found that the cues produced a significant modulation of the SSVEP during evidence presentation. While motor preparation adjustments were observed in the earliest trials, the sensory-level effects only emerged with extended task exposure. Our results suggest that, in addition to strategic adjustments to the decision process, probabilistic information can also induce subtle biases in the encoding of the evidence itself.
<<<
翻译
340.
林海onrush
(2024-04-02 00:39):
#paper, Curriculum Learning and Imitation Learning for Model-free Control on Financial Time-series, doi:https://doi.org/10.48550/arXiv.2311.13326,这篇论文针对金融时间序列的无模型控制问题,提出了一种新颖的解决思路。传统的强化学习方法在这一领域面临训练数据有限且噪声大的挑战。为此,本文探索了将课程学习和模仿学习这两种在机器人领域已有成功应用的范式引入到金融问题中。通过在两个代表性的数据集上的大量实证实验,论文发现课程学习能够显著提升强化学习算法在复杂金融时间序列决策中的表现,优于所有baseline方法。课程学习通过数据增强逐步提高训练任务的难度,体现了 "由易到难" 的学习策略。实验表明,这种适度的数据平滑可以有效降低数据中的噪声,使得强化学习算法更好地捕捉到真实的市场信号。
相比之下,直接应用模仿学习的效果并不理想。进一步的分析表明,这可能是由于模仿学习在去除噪声的同时,也丢失了部分关键的市场信号。从统计学的角度看,模仿学习实现了噪声和信号的分解,但过度的去噪反而损害了策略学习的效果。
本文的理论贡献在于提出了一个信号噪声分解的统计框架,用于解释课程学习和模仿学习在金融时间序列问题上的效果差异。这一框架也为算法的改进提供了新的思路。此外,论文还讨论了一些有待未来进一步探索的方向,包括考察信号噪声分解的非平稳特性,探索其他形式的数据平滑方法,以及将课程学习拓展应用到其他类型的高噪声时间序列学习任务中。
arXiv,
2023.
DOI: 10.48550/arXiv.2311.13326
Abstract:
Curriculum learning and imitation learning have been leveraged extensively inthe robotics domain. However, minimal research has been done on leveragingthese ideas on control tasks over highly stochastic time-series data. Here, …
>>>
Curriculum learning and imitation learning have been leveraged extensively inthe robotics domain. However, minimal research has been done on leveragingthese ideas on control tasks over highly stochastic time-series data. Here, wetheoretically and empirically explore these approaches in a representativecontrol task over complex time-series data. We implement the fundamental ideasof curriculum learning via data augmentation, while imitation learning isimplemented via policy distillation from an oracle. Our findings reveal thatcurriculum learning should be considered a novel direction in improvingcontrol-task performance over complex time-series. Our ample random-seedout-sample empirics and ablation studies are highly encouraging for curriculumlearning for time-series control. These findings are especially encouraging aswe tune all overlapping hyperparameters on the baseline -- giving an advantageto the baseline. On the other hand, we find that imitation learning should beused with caution.
<<<
翻译