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. <<<

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 <<<

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. <<<

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. <<<

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. <<<

Sachiko Yamashita, Masakazu Tanaka, Chieri Ida, Kenichi Kouyama, Setsu Nakae, Taisuke Matsuki, Masataka Tsuda, Tsuyoshi Shirai, Kazuo Kamemura, Yoshisuke Nishi, Joel Moss, Masanao Miwa <<<

Protein targets of polyADP-ribosylation undergo covalent modification with high-molecular-weight, branched poly(ADP-ribose) (PAR) of lengths up to 200 or more ADP-ribose residues derived from NAD. PAR polymerase 1 (PARP1) is the most abundant and well-characterized enzyme involved in PAR biosynthesis. Extensive studies have been carried out to determine how polyADP-ribosylation (PARylation) regulates cell proliferation during cell cycle, with conflicting conclusions. Since significant activation of PARP1 occurs during cell lysis in vitro, we changed the standard method for cell lysis, and using our sensitive ELISA system, quantified without addition of a PAR glycohydrolase inhibitor and clarified that the PAR level is significantly higher in S phase than that in G1. Under normal condition in the absence of exogenous DNA-damaging agent, PAR turns over with a half-life of <40 s; consistent with significant decrease of NAD levels in S phase, which is rescued by PARP inhibitors, in line with the observed rapid turnover of PAR. PARP inhibitors delayed cell cycle in S phase and decreased cell proliferation. Our results underscore the importance of a suitable assay system to measure rapid PAR chain dynamics in living cells and aid our understanding of the function of PARylation during the cell cycle. <<<

Genome editing has been a transformative force in the life sciences and human medicine, offering unprecedented opportunities to dissect complex biological processes and treat the underlying causes of many genetic diseases. CRISPR-based technologies, with their remarkable efficiency and easy programmability, stand at the forefront of this revolution. In this Review, we discuss the current state of CRISPR gene editing technologies in both research and therapy, highlighting limitations that constrain them and the technological innovations that have been developed in recent years to address them. Additionally, we examine and summarize the current landscape of gene editing applications in the context of human health and therapeutics. Finally, we outline potential future developments that could shape gene editing technologies and their applications in the coming years. <<<

Autonomous systems are soon to be ubiquitous, from manufacturing autonomy toagricultural field robots, and from health care assistants to the entertainmentindustry. The majority of these systems are developed with modularsub-components for decision-making, planning, and control that may behand-engineered or learning-based. While these existing approaches have beenshown to perform well under the situations they were specifically designed for,they can perform especially poorly in rare, out-of-distribution scenarios thatwill undoubtedly arise at test-time. The rise of foundation models trained onmultiple tasks with impressively large datasets from a variety of fields hasled researchers to believe that these models may provide common sense reasoningthat existing planners are missing. Researchers posit that this common sensereasoning will bridge the gap between algorithm development and deployment toout-of-distribution tasks, like how humans adapt to unexpected scenarios. Largelanguage models have already penetrated the robotics and autonomous systemsdomains as researchers are scrambling to showcase their potential use cases indeployment. While this application direction is very promising empirically,foundation models are known to hallucinate and generate decisions that maysound reasonable, but are in fact poor. We argue there is a need to step backand simultaneously design systems that can quantify the certainty of a model'sdecision, and detect when it may be hallucinating. In this work, we discuss thecurrent use cases of foundation models for decision-making tasks, provide ageneral definition for hallucinations with examples, discuss existingapproaches to hallucination detection and mitigation with a focus on decisionproblems, and explore areas for further research in this exciting field. <<<

The recent development of experimental methods for measuring chromatin state at single-cell resolution has created a need for computational tools capable of analyzing these datasets. Here we developed Signac, a comprehensive toolkit for the analysis of single-cell chromatin data. Signac enables an end-to-end analysis of single-cell chromatin data, including peak calling, quantification, quality control, dimension reduction, clustering, integration with single-cell gene expression datasets, DNA motif analysis and interactive visualization. Through its seamless compatibility with the Seurat package, Signac facilitates the analysis of diverse multimodal single-cell chromatin data, including datasets that co-assay DNA accessibility with gene expression, protein abundance and mitochondrial genotype. We demonstrate scaling of the Signac framework to analyze datasets containing over 700,000 cells. <<<

OBJECTIVE: To investigate the pregnancy outcomes in a cohort of women who failed to obtain a result in non-invasive prenatal testing (NIPT).DESIGN: Historical cohort study.SETTING: A multicentre private practice in Sydney, Australia.POPULATION: Women who failed to obtain a result from NIPT (n = 131).METHODS: The maternal characteristics, antenatal investigations and pregnancy outcomes for these women were compared with those who obtained a result at the same practice and to the general Australian obstetric population.MAIN OUTCOME MEASURES: Antenatal investigations: pregnancy-associated plasma protein-A (PAPP-A), free β-human chorionic gonadotrophin (β-hCG), placental growth factor (PlGF), uterine artery pulsatility index (PI), mean arterial pressure (MAP). Pregnancy outcomes: chromosomal abnormality, pre-eclampsia, gestational diabetes, small-for-gestational-age (SGA), preterm delivery.RESULTS: Only 1.1% of NIPT samples failed to return a result. This cohort was significantly older and had significantly increased weight compared with the general Australian obstetric population. Pregnancy outcomes were available for 94% of the cohort. There were significantly higher rates of chromosomal aneuploidies (6.5% versus 0.2%, P < 0.0001), pre-eclampsia (11% versus 1.5%, P < 0.0001) and gestational diabetes (23% versus 7.5%, P < 0.0001) compared with the general obstetric population. Rates of preterm delivery and SGA were elevated but did not reach significance. Antenatal investigations demonstrated decreased PAPP-A MoM (0.75 versus 1.14, P < 0.0001), decreased free β-hCG (0.71 versus 1.01, P < 0.0001) and increased uterine artery PI (1.79 versus 1.65, P = 0.02).CONCLUSION: Women who fail to obtain a result from NIPT are at increased risk of adverse pregnancy outcomes, in particular chromosomal aneuploidy, gestational diabetes and pre-eclampsia.FUNDING: None received.TWEETABLE ABSTRACT: Women who fail to obtain a result from cell-free DNA NIPT are at increased risk of adverse pregnancy outcomes. <<<

Xuelei Lin, Yongxin Xu, Dongzhi Wang, Yiman Yang, Xiaoyu Zhang, Xiaomin Bie, Lixuan Gui, Zhongxu Chen, Yiliang Ding, Long Mao, Xueyong Zhang, Fei Lu, Xiansheng Zhang, Cristobal Uauy, Xiangdong Fu, Jun Xiao <<<

The spike architecture of wheat plays a crucial role in determining grain number, making it a key trait for optimization in wheat breeding programs. In this study, we used a multi-omic approach to analyze the transcriptome and epigenome profiles of the young spike at eight developmental stages, revealing coordinated changes in chromatin accessibility and H3K27me3 abundance during the flowering transition. We constructed a core transcriptional regulatory network (TRN) that drives wheat spike formation and experimentally validated a multi-layer regulatory module involving TaSPL15, TaAGLG1, and TaFUL2. By integrating the TRN with genome-wide association studies, we identified 227 transcription factors, including 42 with known functions and 185 with unknown functions. Further investigation of 61 novel transcription factors using multiple homozygous mutant lines revealed 36 transcription factors that regulate spike architecture or flowering time, such as TaMYC2-A1, TaMYB30-A1, and TaWRKY37-A1. Of particular interest, TaMYB30-A1, downstream of and repressed by WFZP, was found to regulate fertile spikelet number. Notably, the excellent haplotype of TaMYB30-A1, which contains a C allele at the WFZP binding site, was enriched during wheat breeding improvement in China, leading to improved agronomic traits. Finally, we constructed a free and open access Wheat Spike Multi-Omic Database (http://39.98.48.156:8800/#/). Our study identifies novel and high-confidence regulators and offers an effective strategy for dissecting the genetic basis of wheat spike development, with practical value for wheat breeding. <<<

Zizhen Yao, Cindy T J van Velthoven, Thuc Nghi Nguyen, Jeff Goldy, Adriana E Sedeno-Cortes, Fahimeh Baftizadeh, Darren Bertagnolli, Tamara Casper, Megan Chiang, Kirsten Crichton, Song-Lin Ding, Olivia Fong, Emma Garren, Alexandra Glandon, Nathan W Gouwens, James Gray, Lucas T Graybuck, Michael J Hawrylycz, Daniel Hirschstein, Matthew Kroll, Kanan Lathia, Changkyu Lee, Boaz Levi, Delissa McMillen, Stephanie Mok, Thanh Pham, Qingzhong Ren, Christine Rimorin, Nadiya Shapovalova, Josef Sulc, Susan M Sunkin, Michael Tieu, Amy Torkelson, Herman Tung, Katelyn Ward, Nick Dee, Kimberly A Smith, Bosiljka Tasic, Hongkui Zeng <<<

The isocortex and hippocampal formation (HPF) in the mammalian brain play critical roles in perception, cognition, emotion, and learning. We profiled ∼1.3 million cells covering the entire adult mouse isocortex and HPF and derived a transcriptomic cell-type taxonomy revealing a comprehensive repertoire of glutamatergic and GABAergic neuron types. Contrary to the traditional view of HPF as having a simpler cellular organization, we discover a complete set of glutamatergic types in HPF homologous to all major subclasses found in the six-layered isocortex, suggesting that HPF and the isocortex share a common circuit organization. We also identify large-scale continuous and graded variations of cell types along isocortical depth, across the isocortical sheet, and in multiple dimensions in hippocampus and subiculum. Overall, our study establishes a molecular architecture of the mammalian isocortex and hippocampal formation and begins to shed light on its underlying relationship with the development, evolution, connectivity, and function of these two brain structures. <<<

Darcy L Fehlings, Mehdi Zarrei, Worrawat Engchuan, Neal Sondheimer, Bhooma Thiruvahindrapuram, Jeffrey R MacDonald, Edward J Higginbotham, Ritesh Thapa, Tarannum Behlim, Sabrina Aimola, Lauren Switzer, Pamela Ng, John Wei, Prakroothi S Danthi, Giovanna Pellecchia, Sylvia Lamoureux, Karen Ho, Sergio L Pereira, Jill de Rijke, Wilson W L Sung, Alireza Mowjoodi, Jennifer L Howe, Thomas Nalpathamkalam, Roozbeh Manshaei, Siavash Ghaffari, Joseph Whitney, Rohan V Patel, Omar Hamdan, Rulan Shaath, Brett Trost, Shannon Knights, Dawa Samdup, Anna McCormick, Carolyn Hunt, Adam Kirton, Anne Kawamura, Ronit Mesterman, Jan Willem Gorter, Nomazulu Dlamini, Daniele Merico, Murto Hilali, Kyle Hirschfeld, Kritika Grover, Nelson X Bautista, Kara Han, Christian R Marshall, Ryan K C Yuen, Padmaja Subbarao, Meghan B Azad, Stuart E Turvey, Piush Mandhane, Theo J Moraes, Elinor Simons, George Maxwell, Michael Shevell, Gregory Costain, Jacques L Michaud, Fadi F Hamdan, Julie Gauthier, Kevin Uguen, Dimitri J Stavropoulos, Richard F Wintle, Maryam Oskoui, Stephen W Scherer <<<

We performed whole-genome sequencing (WGS) in 327 children with cerebral palsy (CP) and their biological parents. We classified 37 of 327 (11.3%) children as having pathogenic/likely pathogenic (P/LP) variants and 58 of 327 (17.7%) as having variants of uncertain significance. Multiple classes of P/LP variants included single-nucleotide variants (SNVs)/indels (6.7%), copy number variations (3.4%) and mitochondrial mutations (1.5%). The COL4A1 gene had the most P/LP SNVs. We also analyzed two pediatric control cohorts (n = 203 trios and n = 89 sib-pair families) to provide a baseline for de novo mutation rates and genetic burden analyses, the latter of which demonstrated associations between de novo deleterious variants and genes related to the nervous system. An enrichment analysis revealed previously undescribed plausible candidate CP genes (SMOC1, KDM5B, BCL11A and CYP51A1). A multifactorial CP risk profile and substantial presence of P/LP variants combine to support WGS in the diagnostic work-up across all CP and related phenotypes. <<<

BACKGROUND: There is divergence in the rate at which people age. The concept of biological age is postulated to capture this variability, and hence to better represent an individual's true global physiological state than chronological age. Biological age predictors are often generated based on cross-sectional data, using biochemical or molecular markers as predictor variables. It is assumed that the difference between chronological and predicted biological age is informative of one's chronological age-independent aging divergence ∆.METHODS: We investigated the statistical assumptions underlying the most popular cross-sectional biological age predictors, based on multiple linear regression, the Klemera-Doubal method or principal component analysis. We used synthetic and real data to illustrate the consequences if this assumption does not hold.RESULTS: The most popular cross-sectional biological age predictors all use the same strong underlying assumption, namely that a candidate marker of aging's association with chronological age is directly informative of its association with the aging rate ∆. We called this the identical-association assumption and proved that it is untestable in a cross-sectional setting. If this assumption does not hold, weights assigned to candidate markers of aging are uninformative, and no more signal may be captured than if markers would have been assigned weights at random.CONCLUSIONS: Cross-sectional methods for predicting biological age commonly use the untestable identical-association assumption, which previous literature in the field had never explicitly acknowledged. These methods have inherent limitations and may provide uninformative results, highlighting the importance of researchers exercising caution in the development and interpretation of cross-sectional biological age predictors. <<<

Chao Zhang, Zihan Xu, Shangda Yang, Guohuan Sun, Lumeng Jia, Zhaofeng Zheng, Quan Gu, Wei Tao, Tao Cheng, Cheng Li, Hui Cheng <<<

Spatiotemporal chromatin reorganization during hematopoietic differentiation has not been comprehensively characterized, mainly because of the large numbers of starting cells required for current chromatin conformation capture approaches. Here, we introduce a low-input tagmentation-based Hi-C (tagHi-C) method to capture the chromatin structures of hundreds of cells. Using tagHi-C, we are able to map the spatiotemporal dynamics of chromatin structure in ten primary hematopoietic stem, progenitor, and differentiated cell populations from mouse bone marrow. Our results reveal that changes in compartment dynamics and the Rabl configuration occur during hematopoietic cell differentiation. We identify gene-body-associating domains (GADs) as general structures for highly expressed genes. Moreover, we extend the body of knowledge regarding genes influenced by genome-wide association study (GWAS) loci through spatial chromatin looping. Our study provides the tagHi-C method for studying the three-dimensional (3D) genome of a small number of cells and maps the comprehensive 3D chromatin landscape of bone marrow hematopoietic cells. <<<