Christina V Theodoris, Ling Xiao, Anant Chopra, Mark D Chaffin, Zeina R Al Sayed, Matthew C Hill, Helene Mantineo, Elizabeth M Brydon, Zexian Zeng, X Shirley Liu, Patrick T Ellinor <<<

Mapping gene networks requires large amounts of transcriptomic data to learn the connections between genes, which impedes discoveries in settings with limited data, including rare diseases and diseases affecting clinically inaccessible tissues. Recently, transfer learning has revolutionized fields such as natural language understanding and computer vision by leveraging deep learning models pretrained on large-scale general datasets that can then be fine-tuned towards a vast array of downstream tasks with limited task-specific data. Here, we developed a context-aware, attention-based deep learning model, Geneformer, pretrained on a large-scale corpus of about 30 million single-cell transcriptomes to enable context-specific predictions in settings with limited data in network biology. During pretraining, Geneformer gained a fundamental understanding of network dynamics, encoding network hierarchy in the attention weights of the model in a completely self-supervised manner. Fine-tuning towards a diverse panel of downstream tasks relevant to chromatin and network dynamics using limited task-specific data demonstrated that Geneformer consistently boosted predictive accuracy. Applied to disease modelling with limited patient data, Geneformer identified candidate therapeutic targets for cardiomyopathy. Overall, Geneformer represents a pretrained deep learning model from which fine-tuning towards a broad range of downstream applications can be pursued to accelerate discovery of key network regulators and candidate therapeutic targets. <<<

Zi-Ning Choo, Julie M Behr, Aditya Deshpande, Kevin Hadi, Xiaotong Yao, Huasong Tian, Kaori Takai, George Zakusilo, Joel Rosiene, Arnaud Da Cruz Paula, Britta Weigelt, Jeremy Setton, Nadeem Riaz, Simon N Powell, Klaus Busam, Alexander N Shoushtari, Charlotte Ariyan, Jorge Reis-Filho, Titia de Lange, Marcin Imieliński <<<

Short-read sequencing is the workhorse of cancer genomics yet is thought to miss many structural variants (SVs), particularly large chromosomal alterations. To characterize missing SVs in short-read whole genomes, we analyzed 'loose ends'-local violations of mass balance between adjacent DNA segments. In the landscape of loose ends across 1,330 high-purity cancer whole genomes, most large (>10-kb) clonal SVs were fully resolved by short reads in the 87% of the human genome where copy number could be reliably measured. Some loose ends represent neotelomeres, which we propose as a hallmark of the alternative lengthening of telomeres phenotype. These pan-cancer findings were confirmed by long-molecule profiles of 38 breast cancer and melanoma cases. Our results indicate that aberrant homologous recombination is unlikely to drive the majority of large cancer SVs. Furthermore, analysis of mass balance in short-read whole genome data provides a surprisingly complete picture of cancer chromosomal structure. <<<

Genome-wide chromatin conformation capture assays provide formidable insights into the spatial organization of genomes. However, due to the complexity of the data structure, their integration in multi-omics workflows remains challenging. We present data structures, computational methods and visualization tools available in Bioconductor to investigate Hi-C, micro-C and other 3C-related data, in R. An online book ( https://bioconductor.org/books/OHCA/ ) further provides prospective end users with a number of workflows to process, import, analyze and visualize any type of chromosome conformation capture data. <<<

Flow-based propagation and spatiotemporal Transformer are two mainstreammechanisms in video inpainting (VI). Despite the effectiveness of thesecomponents, they still suffer from some limitations that affect theirperformance. Previous propagation-based approaches are performed separatelyeither in the image or feature domain. Global image propagation isolated fromlearning may cause spatial misalignment due to inaccurate optical flow.Moreover, memory or computational constraints limit the temporal range offeature propagation and video Transformer, preventing exploration ofcorrespondence information from distant frames. To address these issues, wepropose an improved framework, called ProPainter, which involves enhancedProPagation and an efficient Transformer. Specifically, we introducedual-domain propagation that combines the advantages of image and featurewarping, exploiting global correspondences reliably. We also propose amask-guided sparse video Transformer, which achieves high efficiency bydiscarding unnecessary and redundant tokens. With these components, ProPainteroutperforms prior arts by a large margin of 1.46 dB in PSNR while maintainingappealing efficiency. <<<

Generative pretrained models have achieved remarkable success in various domains such as language and computer vision. Specifically, the combination of large-scale diverse datasets and pretrained transformers has emerged as a promising approach for developing foundation models. Drawing parallels between language and cellular biology (in which texts comprise words; similarly, cells are defined by genes), our study probes the applicability of foundation models to advance cellular biology and genetic research. Using burgeoning single-cell sequencing data, we have constructed a foundation model for single-cell biology, scGPT, based on a generative pretrained transformer across a repository of over 33 million cells. Our findings illustrate that scGPT effectively distills critical biological insights concerning genes and cells. Through further adaptation of transfer learning, scGPT can be optimized to achieve superior performance across diverse downstream applications. This includes tasks such as cell type annotation, multi-batch integration, multi-omic integration, perturbation response prediction and gene network inference. <<<

Xi Cheng, Haroon Popal, Huanqing Wang, Renfen Hu, Yinyin Zang, Mingzhe Zhang, Mark Allen Thornton, Huajian Cai, Yanchao Bi, James Reilly, Ingrid R. Olson, Yin Wang <<<

A defining characteristic of social complexity in Homo sapiens is the diversity of our relationships. We build various types of connections with people in families, schools, workplaces, neighborhoods, and online communities. How do we make sense of such complex systems of human relationships? By using natural language processing, online surveys, laboratory cognitive tasks, and computational modelling on diverse modern cultures across the world (n = 20,425) and ancient cultures across 3,000 years of history, we discovered a universal representational space of relationship concepts, comprised of five principal dimensions (formality, activeness, valence, exchange, equality) and three core categories (hostile, public and private relationships). Our work reveals the fundamental cognitive constructs and cultural principles of relationship knowledge and advances our understanding of human sociality. <<<

Topological Data Analysis is a recent and fast growing field providing a setof new topological and geometric tools to infer relevant features for possiblycomplex data. This paper is a brief introduction, through a few selectedtopics, to basic fundamental and practical aspects of \tda\ for non experts. <<<

Ruichen Li, Haotian Ye, Du Jiang, Xuelan Wen, Chuwei Wang, Zhe Li, Xiang Li, Di He, Ji Chen, Weiluo Ren, Liwei Wang <<<

PURPOSE: To develop and demonstrate the efficacy of a novel head-and-neck multimodality image registration technique using deep-learning-based cross-modality synthesis.METHODS AND MATERIALS: Twenty-five head-and-neck patients received magnetic resonance (MR) and computed tomography (CT) (CTaligned ) scans on the same day with the same immobilization. Fivefold cross validation was used with all of the MR-CT pairs to train a neural network to generate synthetic CTs from MR images. Twenty-four of 25 patients also had a separate CT without immobilization (CTnon-aligned ) and were used for testing. CTnon-aligned 's were deformed to the synthetic CT, and compared to CTnon-aligned registered to MR. The same registrations were performed from MR to CTnon-aligned and from synthetic CT to CTnon-aligned . All registrations used B-splines for modeling the deformation, and mutual information for the objective. Results were evaluated using the 95% Hausdorff distance among spinal cord contours, landmark error, inverse consistency, and Jacobian determinant of the estimated deformation fields.RESULTS: When large initial rigid misalignment is present, registering CT to MRI-derived synthetic CT aligns the cord better than a direct registration. The average landmark error decreased from 9.8 ± 3.1 mm in MR→CTnon-aligned to 6.0 ± 2.1 mm in CTsynth →CTnon-aligned deformable registrations. In the CT to MR direction, the landmark error decreased from 10.0 ± 4.3 mm in CTnon-aligned →MR deformable registrations to 6.6 ± 2.0 mm in CTnon-aligned →CTsynth deformable registrations. The Jacobian determinant had an average value of 0.98. The proposed method also demonstrated improved inverse consistency over the direct method.CONCLUSIONS: We showed that using a deep learning-derived synthetic CT in lieu of an MR for MR→CT and CT→MR deformable registration offers superior results to direct multimodal registration. <<<

Mathias Munschauer, Celina T Nguyen, Klara Sirokman, Christina R Hartigan, Larson Hogstrom, Jesse M Engreitz, Jacob C Ulirsch, Charles P Fulco, Vidya Subramanian, Jenny Chen, Monica Schenone, Mitchell Guttman, Steven A Carr, Eric S Lander <<<

The human genome contains thousands of long non-coding RNAs, but specific biological functions and biochemical mechanisms have been discovered for only about a dozen. A specific long non-coding RNA-non-coding RNA activated by DNA damage (NORAD)-has recently been shown to be required for maintaining genomic stability, but its molecular mechanism is unknown. Here we combine RNA antisense purification and quantitative mass spectrometry to identify proteins that directly interact with NORAD in living cells. We show that NORAD interacts with proteins involved in DNA replication and repair in steady-state cells and localizes to the nucleus upon stimulation with replication stress or DNA damage. In particular, NORAD interacts with RBMX, a component of the DNA-damage response, and contains the strongest RBMX-binding site in the transcriptome. We demonstrate that NORAD controls the ability of RBMX to assemble a ribonucleoprotein complex-which we term NORAD-activated ribonucleoprotein complex 1 (NARC1)-that contains the known suppressors of genomic instability topoisomerase I (TOP1), ALYREF and the PRPF19-CDC5L complex. Cells depleted for NORAD or RBMX display an increased frequency of chromosome segregation defects, reduced replication-fork velocity and altered cell-cycle progression-which represent phenotypes that are mechanistically linked to TOP1 and PRPF19-CDC5L function. Expression of NORAD in trans can rescue defects caused by NORAD depletion, but rescue is significantly impaired when the RBMX-binding site in NORAD is deleted. Our results demonstrate that the interaction between NORAD and RBMX is important for NORAD function, and that NORAD is required for the assembly of the previously unknown topoisomerase complex NARC1, which contributes to maintaining genomic stability. In addition, we uncover a previously unknown function for long non-coding RNAs in modulating the ability of an RNA-binding protein to assemble a higher-order ribonucleoprotein complex. <<<

Temporal binding has been understood as an illusion in timing judgment. When an action triggers an outcome (e.g. a sound) after a brief delay, the action is reported to occur later than if the outcome does not occur, and the outcome is reported to occur earlier than a similar outcome not caused by an action. We show here that an attention mechanism underlies the seeming illusion of timing judgment. In one method, participants watch a rotating clock hand and report event times by noting the clock hand position when the event occurs. We find that visual spatial attention is critically involved in shaping event time reports made in this way. This occurs because action and outcome events result in shifts of attention around the clock rim, thereby biasing the perceived location of the clock hand. Using a probe detection task to measure attention, we show a difference in the distribution of visual spatial attention between a single-event condition (sound only or action only) and a two-event agency condition (action plus sound). Participants accordingly report the timing of the same event (the sound or the action) differently in the two conditions: spatial attentional shifts masquerading as temporal binding. Furthermore, computational modeling based on the attention measure can reproduce the temporal binding effect. Studies that use time judgment as an implicit marker of voluntary agency should first discount the artefactual changes in event timing reports that actually reflect differences in spatial attention. The study also has important implications for related results in mental chronometry obtained with the clock-like method since Wundt, as attention may well be a critical confounding factor in the interpretation of these studies. <<<

Both IL-17A and IL-22 share cellular sources and signaling pathways. They have synergistic action on epithelial cells to stimulate their production of antimicrobial peptides which are protective against infections. However, both interleukins may contribute to ARDS pathology if their production is not controlled. This study aimed to investigate serum levels of IL-17A and IL-22 in relation to the disease outcome in patients with SARS-CoV-2. Serum IL-17A and IL-22 were measured by ELISA in 40 patients with SARS-CoV-2, aged between 2 months and 16 years, (18 had COVID-19 and 22 had multisystem inflammatory syndrome in children "MIS-C") in comparison to 48 age- and sex-matched healthy control children. Patients with COVID-19 and MIS-C had significantly higher serum IL-17A and IL-22 levels than healthy control children (P < 0.001). Increased serum IL-17A and IL-22 levels were found in all patients. Elevated CRP and serum ferritin levels were found in 90% of these patients. Lymphopenia, neutrophilia, neutropenia, thrombocytopenia and elevated ALT, LDH and D-dimer were found in 45%, 42.5 %, 2.5%, 30%, 32.5%, 82.5%, and 65%, respectively of these patients. There were non-significant differences between patients who recovered and those who died or had a residual illness in serum levels of IL-17A, IL-22 and the routine inflammatory markers of COVID-19. In conclusions, serum IL-17A and IL-22 levels were up-regulated in all patients with COVID-19 and MIS-C. Levels of serum IL-17A, IL-22 and the routine inflammatory markers of COVID-19 were not correlated with the disease outcome. Our conclusions are limited by the sample size. <<<

Yanshuo Chu, Enyu Dai, Yating Li, Guangchun Han, Guangsheng Pei, Davis R Ingram, Krupa Thakkar, Jiang-Jiang Qin, Minghao Dang, Xiuning Le, Can Hu, Qing Deng, Ansam Sinjab, Pravesh Gupta, Ruiping Wang, Dapeng Hao, Fuduan Peng, Xinmiao Yan, Yunhe Liu, Shumei Song, Shaojun Zhang, John V Heymach, Alexandre Reuben, Yasir Y Elamin, Melissa P Pizzi, Yang Lu, Rossana Lazcano, Jian Hu, Mingyao Li, Michael Curran, Andrew Futreal, Anirban Maitra, Amir A Jazaeri, Jaffer A Ajani, Charles Swanton, Xiang-Dong Cheng, Hussein A Abbas, Maura Gillison, Krishna Bhat, Alexander J Lazar, Michael Green, Kevin Litchfield, Humam Kadara, Cassian Yee, Linghua Wang <<<

Tumor-infiltrating T cells offer a promising avenue for cancer treatment, yet their states remain to be fully characterized. Here we present a single-cell atlas of T cells from 308,048 transcriptomes across 16 cancer types, uncovering previously undescribed T cell states and heterogeneous subpopulations of follicular helper, regulatory and proliferative T cells. We identified a unique stress response state, T, characterized by heat shock gene expression. T cells are detectable in situ in the tumor microenvironment across various cancer types, mostly within lymphocyte aggregates or potential tertiary lymphoid structures in tumor beds or surrounding tumor edges. T cell states/compositions correlated with genomic, pathological and clinical features in 375 patients from 23 cohorts, including 171 patients who received immune checkpoint blockade therapy. We also found significantly upregulated heat shock gene expression in intratumoral CD4/CD8 cells following immune checkpoint blockade treatment, particularly in nonresponsive tumors, suggesting a potential role of T cells in immunotherapy resistance. Our well-annotated T cell reference maps, web portal and automatic alignment/annotation tool could provide valuable resources for T cell therapy optimization and biomarker discovery. <<<

Proving mathematical theorems at the olympiad level represents a notable milestone in human-level automated reasoning, owing to their reputed difficulty among the world's best talents in pre-university mathematics. Current machine-learning approaches, however, are not applicable to most mathematical domains owing to the high cost of translating human proofs into machine-verifiable format. The problem is even worse for geometry because of its unique translation challenges, resulting in severe scarcity of training data. We propose AlphaGeometry, a theorem prover for Euclidean plane geometry that sidesteps the need for human demonstrations by synthesizing millions of theorems and proofs across different levels of complexity. AlphaGeometry is a neuro-symbolic system that uses a neural language model, trained from scratch on our large-scale synthetic data, to guide a symbolic deduction engine through infinite branching points in challenging problems. On a test set of 30 latest olympiad-level problems, AlphaGeometry solves 25, outperforming the previous best method that only solves ten problems and approaching the performance of an average International Mathematical Olympiad (IMO) gold medallist. Notably, AlphaGeometry produces human-readable proofs, solves all geometry problems in the IMO 2000 and 2015 under human expert evaluation and discovers a generalized version of a translated IMO theorem in 2004. <<<

Cynicism refers to a negative appraisal of human nature-a belief that self-interest is the ultimate motive guiding human behavior. We explored laypersons' beliefs about cynicism and competence and to what extent these beliefs correspond to reality. Four studies showed that laypeople tend to believe in cynical individuals' cognitive superiority. A further three studies based on the data of about 200,000 individuals from 30 countries debunked these lay beliefs as illusionary by revealing that cynical (vs. less cynical) individuals generally do worse on cognitive ability and academic competency tasks. Cross-cultural analyses showed that competent individuals held contingent attitudes and endorsed cynicism only if it was warranted in a given sociocultural environment. Less competent individuals embraced cynicism unconditionally, suggesting that-at low levels of competence-holding a cynical worldview might represent an adaptive default strategy to avoid the potential costs of falling prey to others' cunning. <<<

Francis R. Willett, Erin M. Kunz, Chaofei Fan, Donald T. Avansino, Guy H. Wilson, Eun Young Choi, Foram Kamdar, Matthew F. Glasser, Leigh R. Hochberg, Shaul Druckmann, Krishna V. Shenoy, Jaimie M. Henderson <<<

AbstractSpeech brain–computer interfaces (BCIs) have the potential to restore rapid communication to people with paralysis by decoding neural activity evoked by attempted speech into text1,2 or sound3,4. Early demonstrations, although promising, have not yet achieved accuracies sufficiently high for communication of unconstrained sentences from a large vocabulary1–7. Here we demonstrate a speech-to-text BCI that records spiking activity from intracortical microelectrode arrays. Enabled by these high-resolution recordings, our study participant—who can no longer speak intelligibly owing to amyotrophic lateral sclerosis—achieved a 9.1% word error rate on a 50-word vocabulary (2.7 times fewer errors than the previous state-of-the-art speech BCI2) and a 23.8% word error rate on a 125,000-word vocabulary (the first successful demonstration, to our knowledge, of large-vocabulary decoding). Our participant’s attempted speech was decoded at 62 words per minute, which is 3.4 times as fast as the previous record8 and begins to approach the speed of natural conversation (160 words per minute9). Finally, we highlight two aspects of the neural code for speech that are encouraging for speech BCIs: spatially intermixed tuning to speech articulators that makes accurate decoding possible from only a small region of cortex, and a detailed articulatory representation of phonemes that persists years after paralysis. These results show a feasible path forward for restoring rapid communication to people with paralysis who can no longer speak. <<<

Recent studies found that the deep convolutional neural networks (DCNNs) trained to recognize facial identities spontaneously learned features that support facial expression recognition, and vice versa. Here, we showed that the self-emerged expression-selective units in a VGG-Face trained for facial identification were tuned to distinct basic expressions and, importantly, exhibited hallmarks of human expression recognition (i.e., facial expression confusion and categorical perception). We then investigated whether the emergence of expression-selective units is attributed to either face-specific experience or domain-general processing by conducting the same analysis on a VGG-16 trained for object classification and an untrained VGG-Face without any visual experience, both having the identical architecture with the pretrained VGG-Face. Although similar expression-selective units were found in both DCNNs, they did not exhibit reliable human-like characteristics of facial expression perception. Together, these findings revealed the necessity of domain-specific visual experience of face identity for the development of facial expression perception, highlighting the contribution of nurture to form human-like facial expression perception. <<<

Small extracellular vesicles (sEVs) are emerging as pivotal players in a wide range of physiological and pathological processes. However, a pressing challenge has been the lack of high-throughput techniques capable of unraveling the intricate heterogeneity of sEVs and decoding the underlying cellular behaviors governing sEV secretion. Here we leverage droplet-based single-cell RNA sequencing (scRNA-seq) and introduce an algorithm, SEVtras, to identify sEV-containing droplets and estimate the sEV secretion activity (ESAI) of individual cells. Through extensive validations on both simulated and real datasets, we demonstrate SEVtras' efficacy in capturing sEV-containing droplets and characterizing the secretion activity of specific cell types. By applying SEVtras to four tumor scRNA-seq datasets, we further illustrate that the ESAI can serve as a potent indicator of tumor progression, particularly in the early stages. With the increasing importance and availability of scRNA-seq datasets, SEVtras holds promise in offering valuable extracellular insights into the cell heterogeneity. <<<

Large Multimodal Model (LMM) GPT-4V(ision) endows GPT-4 with visual groundingcapabilities, making it possible to handle certain tasks through the VisualQuestion Answering (VQA) paradigm. This paper explores the potential ofVQA-oriented GPT-4V in the recently popular visual Anomaly Detection (AD) andis the first to conduct qualitative and quantitative evaluations on the popularMVTec AD and VisA datasets. Considering that this task requires bothimage-/pixel-level evaluations, the proposed GPT-4V-AD framework contains threecomponents: 1) Granular Region Division, 2) Prompt Designing, 3)Text2Segmentation for easy quantitative evaluation, and have made somedifferent attempts for comparative analysis. The results show that GPT-4V canachieve certain results in the zero-shot AD task through a VQA paradigm, suchas achieving image-level 77.1/88.0 and pixel-level 68.0/76.6 AU-ROCs on MVTecAD and VisA datasets, respectively. However, its performance still has acertain gap compared to the state-of-the-art zero-shot method, e.g., WinCLIPann CLIP-AD, and further research is needed. This study provides a baselinereference for the research of VQA-oriented LMM in the zero-shot AD task, and wealso post several possible future works. Code is available at\url{https://github.com/zhangzjn/GPT-4V-AD}. <<<

Nils R Winter, Julian Blanke, Ramona Leenings, Jan Ernsting, Lukas Fisch, Kelvin Sarink, Carlotta Barkhau, Daniel Emden, Katharina Thiel, Kira Flinkenflügel, Alexandra Winter, Janik Goltermann, Susanne Meinert, Katharina Dohm, Jonathan Repple, Marius Gruber, Elisabeth J Leehr, Nils Opel, Dominik Grotegerd, Ronny Redlich, Robert Nitsch, Jochen Bauer, Walter Heindel, Joachim Gross, Benjamin Risse, Till F M Andlauer, Andreas J Forstner, Markus M Nöthen, Marcella Rietschel, Stefan G Hofmann, Julia-Katharina Pfarr, Lea Teutenberg, Paula Usemann, Florian Thomas-Odenthal, Adrian Wroblewski, Katharina Brosch, Frederike Stein, Andreas Jansen, Hamidreza Jamalabadi, Nina Alexander, Benjamin Straube, Igor Nenadic, Tilo Kircher, Udo Dannlowski, Tim Hahn <<<

Importance: Biological psychiatry aims to understand mental disorders in terms of altered neurobiological pathways. However, for one of the most prevalent and disabling mental disorders, major depressive disorder (MDD), no informative biomarkers have been identified.Objective: To evaluate whether machine learning (ML) can identify a multivariate biomarker for MDD.Design, Setting, and Participants: This study used data from the Marburg-Münster Affective Disorders Cohort Study, a case-control clinical neuroimaging study. Patients with acute or lifetime MDD and healthy controls aged 18 to 65 years were recruited from primary care and the general population in Münster and Marburg, Germany, from September 11, 2014, to September 26, 2018. The Münster Neuroimaging Cohort (MNC) was used as an independent partial replication sample. Data were analyzed from April 2022 to June 2023.Exposure: Patients with MDD and healthy controls.Main Outcome and Measure: Diagnostic classification accuracy was quantified on an individual level using an extensive ML-based multivariate approach across a comprehensive range of neuroimaging modalities, including structural and functional magnetic resonance imaging and diffusion tensor imaging as well as a polygenic risk score for depression.Results: Of 1801 included participants, 1162 (64.5%) were female, and the mean (SD) age was 36.1 (13.1) years. There were a total of 856 patients with MDD (47.5%) and 945 healthy controls (52.5%). The MNC replication sample included 1198 individuals (362 with MDD [30.1%] and 836 healthy controls [69.9%]). Training and testing a total of 4 million ML models, mean (SD) accuracies for diagnostic classification ranged between 48.1% (3.6%) and 62.0% (4.8%). Integrating neuroimaging modalities and stratifying individuals based on age, sex, treatment, or remission status does not enhance model performance. Findings were replicated within study sites and also observed in structural magnetic resonance imaging within MNC. Under simulated conditions of perfect reliability, performance did not significantly improve. Analyzing model errors suggests that symptom severity could be a potential focus for identifying MDD subgroups.Conclusion and Relevance: Despite the improved predictive capability of multivariate compared with univariate neuroimaging markers, no informative individual-level MDD biomarker-even under extensive ML optimization in a large sample of diagnosed patients-could be identified. <<<