首页 > 最新文献

bioRxiv : the preprint server for biology最新文献

英文 中文
Subregional activity in the dentate gyrus is amplified during elevated cognitive demands. 在认知要求提高时,齿状回的次区域活动会被放大。
Pub Date : 2025-03-06 DOI: 10.1101/2024.10.31.621367
Charlotte Castillon, Shintaro Otsuka, John Armstrong, Anis Contractor

Neural activity in the dentate gyrus (DG) is required for the detection and discrimination of novelty, context and patterns, amongst other cognitive processes. Prior work has demonstrated that there are differences in the activation of granule neurons in the supra and infrapyramidal blades of the DG during a range of hippocampal dependent tasks. Here we used an automated touch screen pattern separation task combined to temporally controlled tagging of active neurons to determine how performance in a cognitively demanding task affected patterns of neural activity in the DG. We found an increase in the blade-biased activity of suprapyramidal mature granule cells (mGCs) during the performance of a high cognitive demand segment of the task, with a further characteristic distribution of active neurons along the apex to blade, and hilar to molecular layer axes. Chemogenetic inhibition of adult-born granule cells (abDGCs) beyond a critical window of their maturation significantly impaired performance of mice during high-demand conditions but not when cognitive demand was low. abDGC inhibition also elevated the total activity of mGCs and disturbed the patterned distribution of active mGCs even in mice that eventually succeeded in the task. Conversely chemogenetic inhibition of mGCs reduced success in the high cognitive demand portion of this task and decreased the global number of active GCs without affecting the patterned distribution of active cells. These findings demonstrate how a high cognitive demand pattern separation task preferentially activates mGCs in subregions of the DG and are consistent with a modulatory role for abDGCs on the dentate circuit which in part governs the spatially organized patterns of activity of mGCs.

齿状回(DG)的神经活动是检测和辨别新奇事物、情境和模式以及其他认知过程所必需的。先前的研究表明,在一系列依赖海马的任务中,齿状回上叶和下叶颗粒神经元的激活存在差异。在这里,我们使用了一种自动触摸屏模式分离任务,并结合对活跃神经元的时间控制标记,来确定在认知要求较高的任务中的表现如何影响 DG 的神经活动模式。我们发现,在执行认知要求较高的任务时,上锥体成熟颗粒细胞(mGCs)的叶片偏向活动增加,活跃神经元沿顶点至叶片轴和腹腔至分子层轴的分布特征进一步增强。在认知需求较高时,化学抑制成体颗粒细胞(abDGCs)超过其成熟的关键窗口期会显著影响小鼠的表现,而在认知需求较低时则不会。abDGCs抑制也会提高mGCs的总活性,并扰乱活性mGCs的分布模式,即使是最终成功完成任务的小鼠也会受到影响。相反,对mGCs的化学抑制则降低了这项任务中高认知需求部分的成功率,并减少了活性GCs的总体数量,但并不影响活性细胞的模式化分布。这些发现证明了高认知需求模式分离任务是如何优先激活 DG 亚区域中的 mGCs 的,这与 abDGCs 在齿状回路中的抑制作用是一致的,而齿状回路在一定程度上支配着 mGCs 活动的空间组织模式。
{"title":"Subregional activity in the dentate gyrus is amplified during elevated cognitive demands.","authors":"Charlotte Castillon, Shintaro Otsuka, John Armstrong, Anis Contractor","doi":"10.1101/2024.10.31.621367","DOIUrl":"10.1101/2024.10.31.621367","url":null,"abstract":"<p><p>Neural activity in the dentate gyrus (DG) is required for the detection and discrimination of novelty, context and patterns, amongst other cognitive processes. Prior work has demonstrated that there are differences in the activation of granule neurons in the supra and infrapyramidal blades of the DG during a range of hippocampal dependent tasks. Here we used an automated touch screen pattern separation task combined to temporally controlled tagging of active neurons to determine how performance in a cognitively demanding task affected patterns of neural activity in the DG. We found an increase in the blade-biased activity of suprapyramidal mature granule cells (mGCs) during the performance of a high cognitive demand segment of the task, with a further characteristic distribution of active neurons along the apex to blade, and hilar to molecular layer axes. Chemogenetic inhibition of adult-born granule cells (abDGCs) beyond a critical window of their maturation significantly impaired performance of mice during high-demand conditions but not when cognitive demand was low. abDGC inhibition also elevated the total activity of mGCs and disturbed the patterned distribution of active mGCs even in mice that eventually succeeded in the task. Conversely chemogenetic inhibition of mGCs reduced success in the high cognitive demand portion of this task and decreased the global number of active GCs without affecting the patterned distribution of active cells. These findings demonstrate how a high cognitive demand pattern separation task preferentially activates mGCs in subregions of the DG and are consistent with a modulatory role for abDGCs on the dentate circuit which in part governs the spatially organized patterns of activity of mGCs.</p>","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11565968/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142650468","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Disentangling value, arousal and valence systems in approach-avoidance behaviors in humans using functional magnetic resonance imaging.
Pub Date : 2025-03-06 DOI: 10.1101/2025.02.19.639143
Dinavahi V P S Murty, Luiz Pessoa

Appetitive and aversive stimuli evoke approach and avoidance behaviors essential for survival and well-being. While affective processing has been extensively examined in terms of arousal and valence, the extent to which value processing is independent from arousal and valence processing in naturalistic contexts remains unclear. We addressed this gap using a naturalistic approach-avoidance task. Ninety-one human participants underwent functional MRI scanning while engaging in approach-avoidance tasks involving two levels of threat (mild or aversive electrical stimulation) and reward (monetary gains). We estimated effect sizes (Cohen's D) across subjects for increasing levels of threat, reward and arousal; for valence (negative vs positive); and for valence-arousal interactions. Effect sizes for threat and reward were strongly positively correlated across brain voxels (r = 0.82), suggesting a strong influence of a shared factor. Spatial independent component analysis decomposed these effect sizes into two independent latent factors, one that represented arousal processing and another that exhibited characteristics of value processing. Importantly, we predicted that valence-arousal interaction effects would increase with latent value effects across voxels, since both valence and arousal contribute to our overall valuation process. We indeed found this to be true. Furthermore, sizable latent value effects were observed in dorsolateral prefrontal cortex, fusiform gyrus and middle temporal gyrus, areas also involved in attention and executive control. Thus, our findings revealed a value system in the human brain that could operate independently of arousal and valence systems during naturalistic approach-avoidance behaviors, providing new insights into the neural mechanisms of affective processing.

{"title":"Disentangling value, arousal and valence systems in approach-avoidance behaviors in humans using functional magnetic resonance imaging.","authors":"Dinavahi V P S Murty, Luiz Pessoa","doi":"10.1101/2025.02.19.639143","DOIUrl":"10.1101/2025.02.19.639143","url":null,"abstract":"<p><p>Appetitive and aversive stimuli evoke approach and avoidance behaviors essential for survival and well-being. While affective processing has been extensively examined in terms of arousal and valence, the extent to which value processing is independent from arousal and valence processing in naturalistic contexts remains unclear. We addressed this gap using a naturalistic approach-avoidance task. Ninety-one human participants underwent functional MRI scanning while engaging in approach-avoidance tasks involving two levels of threat (mild or aversive electrical stimulation) and reward (monetary gains). We estimated effect sizes (Cohen's D) across subjects for increasing levels of threat, reward and arousal; for valence (negative vs positive); and for valence-arousal interactions. Effect sizes for threat and reward were strongly positively correlated across brain voxels (r = 0.82), suggesting a strong influence of a shared factor. Spatial independent component analysis decomposed these effect sizes into two independent latent factors, one that represented arousal processing and another that exhibited characteristics of value processing. Importantly, we predicted that valence-arousal interaction effects would increase with latent value effects across voxels, since both valence and arousal contribute to our overall valuation process. We indeed found this to be true. Furthermore, sizable latent value effects were observed in dorsolateral prefrontal cortex, fusiform gyrus and middle temporal gyrus, areas also involved in attention and executive control. Thus, our findings revealed a value system in the human brain that could operate independently of arousal and valence systems during naturalistic approach-avoidance behaviors, providing new insights into the neural mechanisms of affective processing.</p>","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11870550/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143545551","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Evolution of lateralized gustation in nematodes. 线虫侧向味觉的进化。
Pub Date : 2025-03-06 DOI: 10.1101/2024.08.31.610597
Marisa Mackie, Vivian Vy Le, Heather R Carstensen, Nicole R Kushnir, Dylan L Castro, Ivan M Dimov, Kathleen T Quach, Steven J Cook, Oliver Hobert, Sreekanth H Chalasani, Ray L Hong

Animals with small nervous systems have a limited number of sensory neurons that must encode information from a changing environment. This problem is particularly exacerbated in nematodes that populate a wide variety of distinct ecological niches but only have a few sensory neurons available to encode multiple modalities. How does sensory diversity prevail within this constraint in neuron number? To identify the genetic basis for patterning different nervous systems, we demonstrate that sensory neurons in Pristionchus pacificus respond to various salt sensory cues in a manner that is partially distinct from that of the distantly related nematode Caenorhabditis elegans. By visualizing neuronal activity patterns, we show that contrary to previous expectations based on its genome sequence, the salt responses of P. pacificus are encoded in a left/right asymmetric manner in the bilateral ASE neuron pair. Our study illustrates patterns of evolutionary stability and change in the gustatory system of nematodes.

神经系统较小的动物的感觉神经元数量有限,必须对不断变化的环境信息进行编码。这个问题在线虫中尤为严重,因为线虫栖息在各种不同的生态位中,但只有少数感觉神经元可以编码多种模式的信息。在这种神经元限制条件下,感觉多样性是如何实现的?为了确定不同神经系统模式化的遗传基础,我们证明了太平洋栉水母(Pristionchus pacificus)的感觉神经元对各种盐感觉线索的反应方式与远亲线虫秀丽隐杆线虫(C. elegans)的反应方式部分不同。通过可视化神经元活动模式,我们发现与之前基于其基因组序列的预期相反,太平洋栉水母的盐反应是以左右不对称的方式在双侧 ASE 神经元对中编码的。我们的研究说明了线虫味觉系统的进化稳定性和变化模式。
{"title":"Evolution of lateralized gustation in nematodes.","authors":"Marisa Mackie, Vivian Vy Le, Heather R Carstensen, Nicole R Kushnir, Dylan L Castro, Ivan M Dimov, Kathleen T Quach, Steven J Cook, Oliver Hobert, Sreekanth H Chalasani, Ray L Hong","doi":"10.1101/2024.08.31.610597","DOIUrl":"10.1101/2024.08.31.610597","url":null,"abstract":"<p><p>Animals with small nervous systems have a limited number of sensory neurons that must encode information from a changing environment. This problem is particularly exacerbated in nematodes that populate a wide variety of distinct ecological niches but only have a few sensory neurons available to encode multiple modalities. How does sensory diversity prevail within this constraint in neuron number? To identify the genetic basis for patterning different nervous systems, we demonstrate that sensory neurons in <i>Pristionchus pacificus</i> respond to various salt sensory cues in a manner that is partially distinct from that of the distantly related nematode <i>Caenorhabditis elegans</i>. By visualizing neuronal activity patterns, we show that contrary to previous expectations based on its genome sequence, the salt responses of <i>P. pacificus</i> are encoded in a left/right asymmetric manner in the bilateral ASE neuron pair. Our study illustrates patterns of evolutionary stability and change in the gustatory system of nematodes.</p>","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11398344/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142305877","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
RNAprecis: Prediction of full-detail RNA conformation from the experimentally best-observed sparse parameters.
Pub Date : 2025-03-05 DOI: 10.1101/2025.02.06.636803
Henrik Wiechers, Christopher J Williams, Benjamin Eltzner, Franziska Hoppe, Michael G Prisant, Vincent B Chen, Ezra Miller, Kanti V Mardia, Jane S Richardson, Stephan F Huckemann

We address the problem of predicting high detail RNA structure geometry from the information available in low resolution experimental maps of electron density. Here low resolution refers to ≥2.5Å where the location of the phosphate groups and the glyocosidic bonds can be determined from electron density but all other backbone atom positions cannot. In contrast, high resolution determines all backbone atomic positions. To this end, we firstly create a gold standard data base for four groups of manually corrected suites, each reflecting one out of four sugar pucker-pair configurations. Secondly we develop and employ a modified version of the previously devised algorithm MINT-AGE to learn clusters that are in high correspondence with gold standard's conformational classes based on 3D RNA structure. Since some of the manually corrected classes are of very small size, the modified version of MINT-AGE is able to also identify very small clusters. Thirdly, the new algorithm RNAprecis assigns low resolution structures to newly designed 3D shape coordinates. Our improvements include: (i) learned classes augmented to cover also very low sample sizes and (ii) regularizing a key distance by introducing an adaptive Mahalanobis distance. On a test data containing many clashing and suites modeled as conformational outliers, RNA precis shows good results suggesting that our learning method generalizes well. In particular, our modified MINT-AGE clustering can be finer than the existing curated gold standard suite conformers. For example, the 0a conformer has been separated into two clusters seen in different structural contexts. Such new distinctions can have implications for biochemical interpretation of RNA structure.

我们解决了从低分辨率电子密度实验图中获得的信息预测高细节 RNA 结构几何的问题。这里的低分辨率是指⪆ 2.5埃,其中磷酸基团和糖苷键的位置可以通过电子密度确定,但所有其他骨架原子的位置则无法确定。相反,高分辨率可以确定所有骨架原子的位置。为此,我们首先为四组人工校正套件创建了一个金标准数据库,每组反映四种糖折叠对构型中的一种。其次,我们开发并使用先前设计的 MINT-AGE 算法的改进版,以三维 RNA 结构为基础,学习与金标准构象类别高度对应的聚类。由于一些人工校正的类别非常小,因此改进版 MINT-AGE 也能识别非常小的簇。第三,新算法 RNAprecis 将低分辨率结构分配到新设计的三维形状坐标上。我们的改进包括(i) 对学习到的类别进行了扩充,使其也能覆盖非常小的样本量;(ii) 通过引入自适应的 Mahalanobis 距离对关键距离进行正则化。在包含许多冲突和被建模为构象异常值的套件的测试数据上,RNA precis 显示出良好的结果,表明我们的学习方法具有良好的通用性。特别是,我们修改后的 MINT-AGE 聚类比现有的黄金标准套组构象更精细。例如,0a 构象被分为两个在不同结构背景下出现的聚类。这种新的区分可能会对 RNA 结构的生化解释产生影响:用于生成本文所介绍的分析和图表的代码可在 https://gitlab.gwdg.de/henrik.wiechers1/rnaprecis 上找到。
{"title":"RNAprecis: Prediction of full-detail RNA conformation from the experimentally best-observed sparse parameters.","authors":"Henrik Wiechers, Christopher J Williams, Benjamin Eltzner, Franziska Hoppe, Michael G Prisant, Vincent B Chen, Ezra Miller, Kanti V Mardia, Jane S Richardson, Stephan F Huckemann","doi":"10.1101/2025.02.06.636803","DOIUrl":"10.1101/2025.02.06.636803","url":null,"abstract":"<p><p>We address the problem of predicting high detail RNA structure geometry from the information available in low resolution experimental maps of electron density. Here low resolution refers to ≥2.5Å where the location of the phosphate groups and the glyocosidic bonds can be determined from electron density but all other backbone atom positions cannot. In contrast, high resolution determines all backbone atomic positions. To this end, we firstly create a gold standard data base for four groups of manually corrected suites, each reflecting one out of four sugar pucker-pair configurations. Secondly we develop and employ a modified version of the previously devised algorithm MINT-AGE to learn clusters that are in high correspondence with gold standard's conformational classes based on 3D RNA structure. Since some of the manually corrected classes are of very small size, the modified version of MINT-AGE is able to also identify very small clusters. Thirdly, the new algorithm RNAprecis assigns low resolution structures to newly designed 3D shape coordinates. Our improvements include: (i) learned classes augmented to cover also very low sample sizes and (ii) regularizing a key distance by introducing an adaptive Mahalanobis distance. On a test data containing many clashing and suites modeled as conformational outliers, RNA precis shows good results suggesting that our learning method generalizes well. In particular, our modified MINT-AGE clustering can be finer than the existing curated gold standard suite conformers. For example, the <b>0a</b> conformer has been separated into two clusters seen in different structural contexts. Such new distinctions can have implications for biochemical interpretation of RNA structure.</p>","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11839040/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143461818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Benchmarking DNA Sequence Models for Causal Regulatory Variant Prediction in Human Genetics.
Pub Date : 2025-03-04 DOI: 10.1101/2025.02.11.637758
Gonzalo Benegas, Gokcen Eraslan, Yun S Song

Machine learning holds immense promise in biology, particularly for the challenging task of identifying causal variants for Mendelian and complex traits. Two primary approaches have emerged for this task: supervised sequence-to-function models trained on functional genomics experimental data and self-supervised DNA language models that learn evolutionary constraints on sequences. However, the field currently lacks consistently curated datasets with accurate labels, especially for non-coding variants, that are necessary to comprehensively benchmark these models and advance the field. In this work, we present TraitGym, a curated dataset of regulatory genetic variants that are either known to be causal or are strong candidates across 113 Mendelian and 83 complex traits, along with carefully constructed control variants. We frame the causal variant prediction task as a binary classification problem and benchmark various models, including functional-genomics-supervised models, self-supervised models, models that combine machine learning predictions with curated annotation features, and ensembles of these. Our results provide insights into the capabilities and limitations of different approaches for predicting the functional consequences of non-coding genetic variants. We find that alignment-based models CADD and GPN-MSA compare favorably for Mendelian traits and complex disease traits, while functional-genomics-supervised models Enformer and Borzoi perform better for complex non-disease traits. Evo2 shows substantial performance gains with scale, but still lags somewhat behind alignment-based models, struggling particularly with enhancer variants. The benchmark, including a Google Colab notebook to evaluate a model in a few minutes, is available at https://huggingface.co/datasets/songlab/TraitGym.

机器学习在生物学领域大有可为,特别是对于识别孟德尔性状和复杂性状的因果变异这一具有挑战性的任务。针对这一任务出现了两种主要方法:根据功能基因组学实验数据训练的监督序列到功能模型,以及学习序列进化约束的自监督 DNA 语言模型。然而,该领域目前缺乏具有准确标签(尤其是非编码变异)的一致性数据集,而这些数据集是全面衡量这些模型并推动该领域发展所必需的。在这项工作中,我们展示了 TraitGym,这是一个由调控遗传变异组成的数据集,这些变异要么是已知的因果关系,要么是 113 个孟德尔性状和 83 个复杂性状的有力候选变异,以及精心构建的对照变异。我们将因果变异预测任务视为二元分类问题,并对各种模型进行了基准测试,包括功能基因组学监督模型、自我监督模型、将机器学习预测与标注特征相结合的模型以及这些模型的集合。我们的研究结果让人们深入了解了不同方法预测非编码基因变异功能后果的能力和局限性。我们发现,基于配准的模型 CADD 和 GPN-MSA 在孟德尔性状和复杂疾病性状方面表现较好,而功能基因组学监督模型 Enformer 和 Borzoi 在复杂非疾病性状方面表现更好。该基准包括一个谷歌 Colab 笔记本,可在几分钟内对模型进行评估,可在 https://huggingface.co/datasets/songlab/TraitGym 上查阅。
{"title":"Benchmarking DNA Sequence Models for Causal Regulatory Variant Prediction in Human Genetics.","authors":"Gonzalo Benegas, Gokcen Eraslan, Yun S Song","doi":"10.1101/2025.02.11.637758","DOIUrl":"10.1101/2025.02.11.637758","url":null,"abstract":"<p><p>Machine learning holds immense promise in biology, particularly for the challenging task of identifying causal variants for Mendelian and complex traits. Two primary approaches have emerged for this task: supervised sequence-to-function models trained on functional genomics experimental data and self-supervised DNA language models that learn evolutionary constraints on sequences. However, the field currently lacks consistently curated datasets with accurate labels, especially for non-coding variants, that are necessary to comprehensively benchmark these models and advance the field. In this work, we present TraitGym, a curated dataset of regulatory genetic variants that are either known to be causal or are strong candidates across 113 Mendelian and 83 complex traits, along with carefully constructed control variants. We frame the causal variant prediction task as a binary classification problem and benchmark various models, including functional-genomics-supervised models, self-supervised models, models that combine machine learning predictions with curated annotation features, and ensembles of these. Our results provide insights into the capabilities and limitations of different approaches for predicting the functional consequences of non-coding genetic variants. We find that alignment-based models CADD and GPN-MSA compare favorably for Mendelian traits and complex disease traits, while functional-genomics-supervised models Enformer and Borzoi perform better for complex non-disease traits. Evo2 shows substantial performance gains with scale, but still lags somewhat behind alignment-based models, struggling particularly with enhancer variants. The benchmark, including a Google Colab notebook to evaluate a model in a few minutes, is available at https://huggingface.co/datasets/songlab/TraitGym.</p>","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11844472/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143485305","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Metapipeline-DNA: A Comprehensive Germline & Somatic Genomics Nextflow Pipeline. Metapipeline-DNA:全面的种系和体细胞基因组学 Nextflow 管道。
Pub Date : 2025-03-04 DOI: 10.1101/2024.09.04.611267
Yash Patel, Chenghao Zhu, Takafumi N Yamaguchi, Nicholas K Wang, Nicholas Wiltsie, Nicole Zeltser, Alfredo E Gonzalez, Helena K Winata, Yu Pan, Mohammed Faizal Eeman Mootor, Timothy Sanders, Sorel T Fitz-Gibbon, Cyriac Kandoth, Julie Livingstone, Lydia Y Liu, Benjamin Carlin, Aaron Holmes, Jieun Oh, John Sahrmann, Shu Tao, Stefan Eng, Rupert Hugh-White, Kiarod Pashminehazar, Andrew Park, Arpi Beshlikyan, Madison Jordan, Selina Wu, Mao Tian, Jaron Arbet, Beth Neilsen, Roni Haas, Yuan Zhe Bugh, Gina Kim, Joseph Salmingo, Wenshu Zhang, Aakarsh Anand, Edward Hwang, Anna Neiman-Golden, Philippa Steinberg, Wenyan Zhao, Prateek Anand, Raag Agrawal, Brandon L Tsai, Paul C Boutros

The price, quality and throughout of DNA sequencing continue to improve. Algorithmic innovations have allowed inference of a growing range of features from DNA sequencing data, quantifying nuclear, mitochondrial and evolutionary aspects of both germline and somatic genomes. To automate analyses of the full range of genomic characteristics, we created an extensible Nextflow meta-pipeline called metapipeline-DNA. Metapipeline-DNA analyzes targeted and whole-genome sequencing data from raw reads through pre-processing, feature detection by multiple algorithms, quality-control and data-visualization. Each step can be run independently and is supported robust software engineering including automated failure-recovery, robust testing and consistent verifications of inputs, outputs and parameters. Metapipeline-DNA is cloud-compatible and highly configurable, with options to subset and optimize each analysis. Metapipeline-DNA facilitates high-scale, comprehensive analysis of DNA sequencing data.

摘要:随着高通量技术的发展,DNA 测序的价格越来越低,速度越来越快。数据可用性的提高促进了新型算法的开发,以阐明以前模糊不清的特征,并导致人们越来越依赖复杂的工作流程,将这些工具集成到分析管道中。为了便于分析 DNA 测序数据,我们创建了 metapipeline-DNA,这是一个高度可配置和可扩展的管道。它涵盖了广泛的处理过程,包括原始测序读数比对和重新校准、变体调用、质量控制和亚克隆重建。Metapipeline-DNA 还包含配置选项,用于选择和调整分析,同时对故障具有鲁棒性。这使得在临床和研究环境中分析大型 DNA 测序的能力标准化和简单化:Metapipeline-DNA 是开源的 Nextflow 管道,采用 GPLv2 许可,可在 https://github.com/uclahs-cds/metapipeline-DNA 免费获取。
{"title":"Metapipeline-DNA: A Comprehensive Germline & Somatic Genomics Nextflow Pipeline.","authors":"Yash Patel, Chenghao Zhu, Takafumi N Yamaguchi, Nicholas K Wang, Nicholas Wiltsie, Nicole Zeltser, Alfredo E Gonzalez, Helena K Winata, Yu Pan, Mohammed Faizal Eeman Mootor, Timothy Sanders, Sorel T Fitz-Gibbon, Cyriac Kandoth, Julie Livingstone, Lydia Y Liu, Benjamin Carlin, Aaron Holmes, Jieun Oh, John Sahrmann, Shu Tao, Stefan Eng, Rupert Hugh-White, Kiarod Pashminehazar, Andrew Park, Arpi Beshlikyan, Madison Jordan, Selina Wu, Mao Tian, Jaron Arbet, Beth Neilsen, Roni Haas, Yuan Zhe Bugh, Gina Kim, Joseph Salmingo, Wenshu Zhang, Aakarsh Anand, Edward Hwang, Anna Neiman-Golden, Philippa Steinberg, Wenyan Zhao, Prateek Anand, Raag Agrawal, Brandon L Tsai, Paul C Boutros","doi":"10.1101/2024.09.04.611267","DOIUrl":"10.1101/2024.09.04.611267","url":null,"abstract":"<p><p>The price, quality and throughout of DNA sequencing continue to improve. Algorithmic innovations have allowed inference of a growing range of features from DNA sequencing data, quantifying nuclear, mitochondrial and evolutionary aspects of both germline and somatic genomes. To automate analyses of the full range of genomic characteristics, we created an extensible Nextflow meta-pipeline called metapipeline-DNA. Metapipeline-DNA analyzes targeted and whole-genome sequencing data from raw reads through pre-processing, feature detection by multiple algorithms, quality-control and data-visualization. Each step can be run independently and is supported robust software engineering including automated failure-recovery, robust testing and consistent verifications of inputs, outputs and parameters. Metapipeline-DNA is cloud-compatible and highly configurable, with options to subset and optimize each analysis. Metapipeline-DNA facilitates high-scale, comprehensive analysis of DNA sequencing data.</p>","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11398472/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142305954","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Spatiotemporal Patterns Differentiate Hippocampal Sharp-Wave Ripples from Interictal Epileptiform Discharges in Mice and Humans.
Pub Date : 2025-03-04 DOI: 10.1101/2025.02.06.636758
Anna Maslarova, Jiyun N Shin, Andrea Navas-Olive, Mihály Vöröslakos, Hajo Hamer, Arnd Doerfler, Simon Henin, György Buzsáki, Anli Liu

Hippocampal sharp-wave ripples (SPW-Rs) are high-frequency oscillations critical for memory consolidation in mammals. Despite extensive characterization in rodents, their application as biomarkers to track and treat memory dysfunction in humans is limited by coarse spatial sampling, interference from interictal epileptiform discharges (IEDs), and lack of consensus on human SPW-R localization and morphology. We demonstrate that mouse and human hippocampal ripples share spatial, spectral and temporal features, which are clearly distinct from IEDs. In 1024-channel hippocampal recordings from APP/PS1 mice, SPW-Rs were distinguishable from IEDs by their narrow localization to the CA1 pyramidal layer, narrowband frequency peaks, and multiple ripple cycles on the unfiltered local field potential. In epilepsy patients, ripples showed similar narrowband frequency peaks and visible ripple cycles in CA1 and the subiculum but were absent in the dentate gyrus. Conversely, IEDs showed a broad spatial extent and wide-band frequency power. We introduce a semi-automated, human ripple detection toolbox ("ripmap") selecting optimal detection channels and separating event waveforms by low-dimensional embedding. Our approach improves ripple detection accuracy, providing a firm foundation for future human memory research.

{"title":"Spatiotemporal Patterns Differentiate Hippocampal Sharp-Wave Ripples from Interictal Epileptiform Discharges in Mice and Humans.","authors":"Anna Maslarova, Jiyun N Shin, Andrea Navas-Olive, Mihály Vöröslakos, Hajo Hamer, Arnd Doerfler, Simon Henin, György Buzsáki, Anli Liu","doi":"10.1101/2025.02.06.636758","DOIUrl":"10.1101/2025.02.06.636758","url":null,"abstract":"<p><p>Hippocampal sharp-wave ripples (SPW-Rs) are high-frequency oscillations critical for memory consolidation in mammals. Despite extensive characterization in rodents, their application as biomarkers to track and treat memory dysfunction in humans is limited by coarse spatial sampling, interference from interictal epileptiform discharges (IEDs), and lack of consensus on human SPW-R localization and morphology. We demonstrate that mouse and human hippocampal ripples share spatial, spectral and temporal features, which are clearly distinct from IEDs. In 1024-channel hippocampal recordings from APP/PS1 mice, SPW-Rs were distinguishable from IEDs by their narrow localization to the CA1 pyramidal layer, narrowband frequency peaks, and multiple ripple cycles on the unfiltered local field potential. In epilepsy patients, ripples showed similar narrowband frequency peaks and visible ripple cycles in CA1 and the subiculum but were absent in the dentate gyrus. Conversely, IEDs showed a broad spatial extent and wide-band frequency power. We introduce a semi-automated, human ripple detection toolbox (\"ripmap\") selecting optimal detection channels and separating event waveforms by low-dimensional embedding. Our approach improves ripple detection accuracy, providing a firm foundation for future human memory research.</p>","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11839046/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143461853","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Neoplastic immune mimicry potentiates breast tumor progression.
Pub Date : 2025-03-04 DOI: 10.1101/2025.01.17.633673
Eric B Berens, Sokchea Khou, Elaine Huang, Amber Hoffman, Briana Johnson, Nell Kirchberger, Shamilene Sivagnanam, Nicholas L Calistri, Daniel Derrick, Tiera A Liby, Ian C McLean, Aryn A Alanizi, Furkan Ozmen, Tugba Y Ozmen, Gordon B Mills, E Shelley Hwang, Pepper J Schedin, Hugo Gonzalez, Zena Werb, Laura M Heiser, Lisa M Coussens

Dedifferentiation programs are commonly enacted during breast cancer progression to enhance tumor cell fitness. Increased cellular plasticity within the neoplastic compartment of tumors correlates with disease aggressiveness, often culminating in greater resistance to cytotoxic therapies or augmented metastatic potential. Here we report that subpopulations of dedifferentiated neoplastic breast epithelial cells express canonical leukocyte cell surface receptor proteins and have thus named this cellular program "immune mimicry." We document neoplastic cells engaging in immune mimicry within public human breast tumor single-cell RNA-seq datasets, histopathological breast tumor specimens, breast cancer cell lines, as well as in murine transgenic and cell line-derived mammary cancer models. Immune-mimicked neoplastic cells harbor hallmarks of dedifferentiation and are enriched in treatment-resistant and high-grade breast tumors. We corroborated these observations in aggressive breast cancer cell lines where anti-proliferative cytotoxic chemotherapies drove epithelial cells toward immune mimicry. Moreover, in subsequent proof-of-concept studies, we demonstrate that expression of the CD69 leukocyte activation protein by neoplastic cells confers a proliferative advantage that facilitates early tumor growth and therefore conclude that neoplastic breast epithelial cells upregulating leukocyte surface receptors potentiate malignancy. Moving forward, neoplastic immune mimicry should be evaluated for prognostic utility in breast cancer to determine stratification potential for patients with increased risks of tumor recurrence, metastasis, and therapeutic resistance.

Statement of significance: Neoplastic breast epithelial cells express surface receptors canonically attributed to leukocytes and are associated with therapy resistance and aggressive tumor behavior.

{"title":"Neoplastic immune mimicry potentiates breast tumor progression.","authors":"Eric B Berens, Sokchea Khou, Elaine Huang, Amber Hoffman, Briana Johnson, Nell Kirchberger, Shamilene Sivagnanam, Nicholas L Calistri, Daniel Derrick, Tiera A Liby, Ian C McLean, Aryn A Alanizi, Furkan Ozmen, Tugba Y Ozmen, Gordon B Mills, E Shelley Hwang, Pepper J Schedin, Hugo Gonzalez, Zena Werb, Laura M Heiser, Lisa M Coussens","doi":"10.1101/2025.01.17.633673","DOIUrl":"10.1101/2025.01.17.633673","url":null,"abstract":"<p><p>Dedifferentiation programs are commonly enacted during breast cancer progression to enhance tumor cell fitness. Increased cellular plasticity within the neoplastic compartment of tumors correlates with disease aggressiveness, often culminating in greater resistance to cytotoxic therapies or augmented metastatic potential. Here we report that subpopulations of dedifferentiated neoplastic breast epithelial cells express canonical leukocyte cell surface receptor proteins and have thus named this cellular program \"immune mimicry.\" We document neoplastic cells engaging in immune mimicry within public human breast tumor single-cell RNA-seq datasets, histopathological breast tumor specimens, breast cancer cell lines, as well as in murine transgenic and cell line-derived mammary cancer models. Immune-mimicked neoplastic cells harbor hallmarks of dedifferentiation and are enriched in treatment-resistant and high-grade breast tumors. We corroborated these observations in aggressive breast cancer cell lines where anti-proliferative cytotoxic chemotherapies drove epithelial cells toward immune mimicry. Moreover, in subsequent proof-of-concept studies, we demonstrate that expression of the CD69 leukocyte activation protein by neoplastic cells confers a proliferative advantage that facilitates early tumor growth and therefore conclude that neoplastic breast epithelial cells upregulating leukocyte surface receptors potentiate malignancy. Moving forward, neoplastic immune mimicry should be evaluated for prognostic utility in breast cancer to determine stratification potential for patients with increased risks of tumor recurrence, metastasis, and therapeutic resistance.</p><p><strong>Statement of significance: </strong>Neoplastic breast epithelial cells express surface receptors canonically attributed to leukocytes and are associated with therapy resistance and aggressive tumor behavior.</p>","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11785120/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143082951","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Preventing peritendinous adhesions using lubricious supramolecular hydrogels.
Pub Date : 2025-03-04 DOI: 10.1101/2025.02.22.639655
Emily L Meany, Christian M Williams, Ye Eun Song, Vanessa M Doulames, Sophia J Bailey, Shoshana C Williams, Carolyn K Jons, Paige M Fox, Eric A Appel

Of the 1.5 million emergency room visits each year in the United States due to flexor tendon injuries in the hand, over 30-40% result in peritendinous adhesions which can limit range of motion (ROM) and severely impact an individual's quality of life. Adhesions are fibrous scar-like tissues which can form between adjacent tissues in the body in response to injury, inflammation, or during normal healing following surgery. Currently, there is no widespread solution for adhesion prevention in the delicate space of the digit while allowing a patient full ROM quickly after surgery. There is a clear clinical need for a material capable of limiting adhesion formation which is simple to apply, does not impair healing, remains at the application site during motion and initial inflammation (days - weeks), and leaves tendon glide unencumbered. In this work, we developed dynamically crosslinked, bioresorbable supramolecular hydrogels as easy-to-apply lubricious barriers to prevent the formation of peritendinous adhesions. These hydrogels exhibit excellent long-term stability, injectability, and thermally stable viscoelastic properties that allow for simple storage and facile application. We evaluated interactions at the interface of the hydrogels and relevant tissues, including human tendon and skin, in shear and extensional stress modes and demonstrated a unique mechanism of adhesion prevention based on maintenance of a lubricious hydrogel barrier between tissues. Ex vivo studies show that the hydrogels did not impair the gliding behavior nor mechanical properties of tendons when applied in cadaveric human hands following clinically relevant flexor tendon repair. We further applied these hydrogels in a preclinical rat Achilles tendon injury model and observed prolonged local retention at the repair site as well as improved recovery of key functional metrics, including ROM and maximal dorsiflexion. Further, these hydrogels were safe and did not impair tendon strength nor healing compared to the current standard of care. These dynamic, biocompatible hydrogels present a novel solution to the significant problem of peritendinous adhesions with clear translational potential.

{"title":"Preventing peritendinous adhesions using lubricious supramolecular hydrogels.","authors":"Emily L Meany, Christian M Williams, Ye Eun Song, Vanessa M Doulames, Sophia J Bailey, Shoshana C Williams, Carolyn K Jons, Paige M Fox, Eric A Appel","doi":"10.1101/2025.02.22.639655","DOIUrl":"10.1101/2025.02.22.639655","url":null,"abstract":"<p><p>Of the 1.5 million emergency room visits each year in the United States due to flexor tendon injuries in the hand, over 30-40% result in peritendinous adhesions which can limit range of motion (ROM) and severely impact an individual's quality of life. Adhesions are fibrous scar-like tissues which can form between adjacent tissues in the body in response to injury, inflammation, or during normal healing following surgery. Currently, there is no widespread solution for adhesion prevention in the delicate space of the digit while allowing a patient full ROM quickly after surgery. There is a clear clinical need for a material capable of limiting adhesion formation which is simple to apply, does not impair healing, remains at the application site during motion and initial inflammation (days - weeks), and leaves tendon glide unencumbered. In this work, we developed dynamically crosslinked, bioresorbable supramolecular hydrogels as easy-to-apply lubricious barriers to prevent the formation of peritendinous adhesions. These hydrogels exhibit excellent long-term stability, injectability, and thermally stable viscoelastic properties that allow for simple storage and facile application. We evaluated interactions at the interface of the hydrogels and relevant tissues, including human tendon and skin, in shear and extensional stress modes and demonstrated a unique mechanism of adhesion prevention based on maintenance of a lubricious hydrogel barrier between tissues. <i>Ex vivo</i> studies show that the hydrogels did not impair the gliding behavior nor mechanical properties of tendons when applied in cadaveric human hands following clinically relevant flexor tendon repair. We further applied these hydrogels in a preclinical rat Achilles tendon injury model and observed prolonged local retention at the repair site as well as improved recovery of key functional metrics, including ROM and maximal dorsiflexion. Further, these hydrogels were safe and did not impair tendon strength nor healing compared to the current standard of care. These dynamic, biocompatible hydrogels present a novel solution to the significant problem of peritendinous adhesions with clear translational potential.</p>","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11888243/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143589468","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
SensitiveCancerGPT: Leveraging Generative Large Language Model on Structured Omics Data to Optimize Drug Sensitivity Prediction.
Pub Date : 2025-03-03 DOI: 10.1101/2025.02.27.640661
Shaika Chowdhury, Sivaraman Rajaganapathy, Lichao Sun, Liewei Wang, Ping Yang, James R Cerhan, Nansu Zong
<p><strong>Objective: </strong>The fast accumulation of vast pharmacogenomics data of cancer cell lines provide unprecedented opportunities for drug sensitivity prediction (DSP), a crucial prerequisite for the advancement of precision oncology. Recently, Generative Large Language Models (LLM) have demonstrated performance and generalization prowess across diverse tasks in the field of natural language processing (NLP). However, the structured format of the pharmacogenomics data poses challenge for the utility of LLM in DSP. Therefore, the objective of this study is multi-fold: to adapt prompt engineering for structured pharmacogenomics data toward optimizing LLM's DSP performance, to evaluate LLM's generalization in real-world DSP scenarios, and to compare LLM's DSP performance against that of state-of-the-science baselines.</p><p><strong>Methods: </strong>We systematically investigated the capability of the Generative Pre-trained Transformer (GPT) as a DSP model on four publicly available benchmark pharmacogenomics datasets, which are stratified by five cancer tissue types of cell lines and encompass both oncology and non-oncology drugs. Essentially, the predictive landscape of GPT is assessed for effectiveness on the DSP task via four learning paradigms: zero-shot learning, few-shot learning, fine-tuning and clustering pretrained embeddings. To facilitate GPT in seamlessly processing the structured pharmacogenomics data, domain-specific novel prompt engineering is employed by implementing three prompt templates (i.e., Instruction, Instruction-Prefix, Cloze) and integrating pharmacogenomics-related features into the prompt. We validated GPT's performance in diverse real-world DSP scenarios: cross-tissue generalization, blind tests, and analyses of drug-pathway associations and top sensitive/resistant cell lines. Furthermore, we conducted a comparative evaluation of GPT against multiple Transformer-based pretrained models and existing DSP baselines.</p><p><strong>Results: </strong>Extensive experiments on the pharmacogenomics datasets across the five tissue cohorts demonstrate that fine-tuning GPT yields the best DSP performance (28% F1 increase, p-value= 0.0003) followed by clustering pretrained GPT embeddings (26% F1 increase, p-value= 0.0005), outperforming GPT in-context learning (i.e., few-shot). However, GPT in the zero-shot setting had a big F1 gap, resulting in the worst performance. Within the scope of prompt engineering, performance enhancement was achieved by directly instructing GPT about the DSP task and resorting to a concise context format (i.e., instruction-prefix), leading to F1 performance gain of 22% (p-value=0.02); while incorporation of drug-cell line prompt context derived from genomics and/or molecular features further boosted F1 score by 2%. Compared to state-of-the-science DSP baselines, GPT significantly asserted superior mean F1 performance (16% gain, p-value<0.05) on the GDSC dataset. In the cross-tissue analysis, GPT
{"title":"SensitiveCancerGPT: Leveraging Generative Large Language Model on Structured Omics Data to Optimize Drug Sensitivity Prediction.","authors":"Shaika Chowdhury, Sivaraman Rajaganapathy, Lichao Sun, Liewei Wang, Ping Yang, James R Cerhan, Nansu Zong","doi":"10.1101/2025.02.27.640661","DOIUrl":"10.1101/2025.02.27.640661","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;The fast accumulation of vast pharmacogenomics data of cancer cell lines provide unprecedented opportunities for drug sensitivity prediction (DSP), a crucial prerequisite for the advancement of precision oncology. Recently, Generative Large Language Models (LLM) have demonstrated performance and generalization prowess across diverse tasks in the field of natural language processing (NLP). However, the structured format of the pharmacogenomics data poses challenge for the utility of LLM in DSP. Therefore, the objective of this study is multi-fold: to adapt prompt engineering for structured pharmacogenomics data toward optimizing LLM's DSP performance, to evaluate LLM's generalization in real-world DSP scenarios, and to compare LLM's DSP performance against that of state-of-the-science baselines.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;We systematically investigated the capability of the Generative Pre-trained Transformer (GPT) as a DSP model on four publicly available benchmark pharmacogenomics datasets, which are stratified by five cancer tissue types of cell lines and encompass both oncology and non-oncology drugs. Essentially, the predictive landscape of GPT is assessed for effectiveness on the DSP task via four learning paradigms: zero-shot learning, few-shot learning, fine-tuning and clustering pretrained embeddings. To facilitate GPT in seamlessly processing the structured pharmacogenomics data, domain-specific novel prompt engineering is employed by implementing three prompt templates (i.e., Instruction, Instruction-Prefix, Cloze) and integrating pharmacogenomics-related features into the prompt. We validated GPT's performance in diverse real-world DSP scenarios: cross-tissue generalization, blind tests, and analyses of drug-pathway associations and top sensitive/resistant cell lines. Furthermore, we conducted a comparative evaluation of GPT against multiple Transformer-based pretrained models and existing DSP baselines.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Extensive experiments on the pharmacogenomics datasets across the five tissue cohorts demonstrate that fine-tuning GPT yields the best DSP performance (28% F1 increase, p-value= 0.0003) followed by clustering pretrained GPT embeddings (26% F1 increase, p-value= 0.0005), outperforming GPT in-context learning (i.e., few-shot). However, GPT in the zero-shot setting had a big F1 gap, resulting in the worst performance. Within the scope of prompt engineering, performance enhancement was achieved by directly instructing GPT about the DSP task and resorting to a concise context format (i.e., instruction-prefix), leading to F1 performance gain of 22% (p-value=0.02); while incorporation of drug-cell line prompt context derived from genomics and/or molecular features further boosted F1 score by 2%. Compared to state-of-the-science DSP baselines, GPT significantly asserted superior mean F1 performance (16% gain, p-value&lt;0.05) on the GDSC dataset. In the cross-tissue analysis, GPT ","PeriodicalId":519960,"journal":{"name":"bioRxiv : the preprint server for biology","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11888479/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143589424","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
期刊
bioRxiv : the preprint server for biology
全部 Acc. Chem. Res. ACS Applied Bio Materials ACS Appl. Electron. Mater. ACS Appl. Energy Mater. ACS Appl. Mater. Interfaces ACS Appl. Nano Mater. ACS Appl. Polym. Mater. ACS BIOMATER-SCI ENG ACS Catal. ACS Cent. Sci. ACS Chem. Biol. ACS Chemical Health & Safety ACS Chem. Neurosci. ACS Comb. Sci. ACS Earth Space Chem. ACS Energy Lett. ACS Infect. Dis. ACS Macro Lett. ACS Mater. Lett. ACS Med. Chem. Lett. ACS Nano ACS Omega ACS Photonics ACS Sens. ACS Sustainable Chem. Eng. ACS Synth. Biol. Anal. Chem. BIOCHEMISTRY-US Bioconjugate Chem. BIOMACROMOLECULES Chem. Res. Toxicol. Chem. Rev. Chem. Mater. CRYST GROWTH DES ENERG FUEL Environ. Sci. Technol. Environ. Sci. Technol. Lett. Eur. J. Inorg. Chem. IND ENG CHEM RES Inorg. Chem. J. Agric. Food. Chem. J. Chem. Eng. Data J. Chem. Educ. J. Chem. Inf. Model. J. Chem. Theory Comput. J. Med. Chem. J. Nat. Prod. J PROTEOME RES J. Am. Chem. Soc. LANGMUIR MACROMOLECULES Mol. Pharmaceutics Nano Lett. Org. Lett. ORG PROCESS RES DEV ORGANOMETALLICS J. Org. Chem. J. Phys. Chem. J. Phys. Chem. A J. Phys. Chem. B J. Phys. Chem. C J. Phys. Chem. Lett. Analyst Anal. Methods Biomater. Sci. Catal. Sci. Technol. Chem. Commun. Chem. Soc. Rev. CHEM EDUC RES PRACT CRYSTENGCOMM Dalton Trans. Energy Environ. Sci. ENVIRON SCI-NANO ENVIRON SCI-PROC IMP ENVIRON SCI-WAT RES Faraday Discuss. Food Funct. Green Chem. Inorg. Chem. Front. Integr. Biol. J. Anal. At. Spectrom. J. Mater. Chem. A J. Mater. Chem. B J. Mater. Chem. C Lab Chip Mater. Chem. Front. Mater. Horiz. MEDCHEMCOMM Metallomics Mol. Biosyst. Mol. Syst. Des. Eng. Nanoscale Nanoscale Horiz. Nat. Prod. Rep. New J. Chem. Org. Biomol. Chem. Org. Chem. Front. PHOTOCH PHOTOBIO SCI PCCP Polym. Chem.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1