Recent publications
Federated learning is a fast-developing distributed learning scheme with promising applications in vertical domains such as industrial automation and connected automated driving. The heterogeneity of devices in data distribution, communication, and computation, when deployed in dynamic environments typically with wireless communication, poses challenges to traditional federated learning solutions, where successful learning depends on balanced contribution from participants. In this paper, we propose a flexible communication strategy for devices in asynchronous federated learning, which adapts the training and uploading actions based on the condition of the communication link. We propose a novel method of computing aggregation weight based on model distances and number of local optimizations, to control errors introduced in asynchronous aggregation while maximizing learning speed. We prove the convergence of the learning tasks analytically under the new scheme. The improved performance is rooted in the increased number of optimizations during training, which grows by 12% through opportunistically condensing model uploading during good link condition periods. By facilitating timely communication between devices and server, combined with the novel aggregation weight design, our method reduces the communication resources in dynamic environments by at least 5% while even slightly increasing the learning accuracy.
- Bernhard Föllmer
- Michelle C. Williams
- Damini Dey
- [...]
- Marc Dewey
Artificial intelligence (AI) is likely to revolutionize the way medical images are analysed and has the potential to improve the identification and analysis of vulnerable or high-risk atherosclerotic plaques in coronary arteries, leading to advances in the treatment of coronary artery disease. However, coronary plaque analysis is challenging owing to cardiac and respiratory motion, as well as the small size of cardiovascular structures. Moreover, the analysis of coronary imaging data is time-consuming, can be performed only by clinicians with dedicated cardiovascular imaging training and is subject to considerable interreader and intrareader variability. AI has the potential to improve the assessment of images of vulnerable plaque in coronary arteries but requires robust development, testing and validation. Combining human expertise with AI might facilitate the reliable and valid interpretation of images obtained using CT, MRI, PET, intravascular ultrasonography and optical coherence tomography. In this roadmap, we review existing evidence on the application of AI to the imaging of vulnerable plaque in coronary arteries and provide consensus recommendations developed by an interdisciplinary group of experts on AI and noninvasive and invasive coronary imaging. We also outline future requirements of AI technology to address bias, uncertainty, explainability and generalizability, which are all essential for the acceptance of AI and its clinical utility in handling the anticipated growing volume of coronary imaging procedures.
Aims
Myocardial inflammation and impaired mitochondrial oxidative capacity are hallmarks of heart failure (HF) pathophysiology. The extent of myocardial inflammation in patients suffering from ischaemic cardiomyopathy (ICM) or dilated cardiomyopathy (DCM) and its association with mitochondrial energy metabolism are unknown. We aimed at establishing a relevant role of cardiac inflammation in the impairment of mitochondrial energy production in advanced ischaemic and non‐ischaemic HF.
Methods
We included 81 patients with stage D HF (ICM, n = 44; DCM, n = 37) undergoing left ventricular assist device implantation (n = 59) or heart transplantation (n = 22) and obtained left ventricular tissue samples during open heart surgery. We quantified mitochondrial oxidative capacity, citrate synthase activity (CSA) and fibrosis and lymphocytic infiltration. We considered infiltration of >14 CD3⁺ cells/mm² relevant inflammation.
Results
Patients with ICM or DCM did not differ regarding age (61.5 ± 5.7 vs. 56.5 ± 12.7 years, P = 0.164), sex (86% vs. 84% male, P = 0.725), type 2 diabetes mellitus (34% vs. 18%, P = 0.126) or chronic kidney disease (8% vs. 11%, P = 0.994). ICM exhibited oxidative capacity reduced by 23% compared to DCM (108.6 ± 41.4 vs. 141.9 ± 59.9 pmol/(s*mg), P = 0.006). Maximum production of reactive oxygen species was not significantly different between ICM and DCM (0.59 ± 0.28 vs. 0.69 ± 0.36 pmol/(s*ml), P = 0.196). Mitochondrial content, detected by CSA, was lower in ICM (359.6 ± 164.1 vs. 503.0 ± 198.5 nmol/min/mg protein, P = 0.002). Notably, relevant inflammation was more common in ICM (27% vs. 6%, P = 0.024), and the absolute number of infiltrating leucocytes correlated with lower oxidative capacity (r = −0.296, P = 0.019). Fibrosis was more prevalent in ICM (20.9 ± 21.2 vs. 7.2 ± 5.6% of area, P = 0.002), but not associated with oxidative capacity (r = −0.13, P = 0.327).
Conclusions
More than every fourth ICM patient with advanced HF displays myocardial inflammation in the range of inflammatory cardiomyopathy associated with reduced mitochondrial oxidative capacity. Future studies may evaluate inflammation in ICM at earlier stages in standardised fashion to explore the therapeutic potential of immunosuppression to influence trajectories of HF in ICM.
Background
Sensitivity to ionizing radiation differs between individuals, but there is a limited understanding of the biological mechanisms that account for these variations. One example of such mechanisms are the mutations in the ATM (mutated ataxia telangiectasia) gene, that cause the rare recessively inherited disease Ataxia telangiectasia (AT). Hallmark features include chromosomal instability and increased sensitivity to ionizing radiation (IR).
Objectives
To deepen the molecular understanding of radiosensitivity and to identify potential new markers to predict it, human ATM-mutated and proficient cells were compared on a proteomic level.
Design
In this study, we analyzed 3 cell lines from AT patients, with varying radiosensitivity, and 2 cell lines from healthy volunteers, 24 hours and 72 hours post-10 Gy irradiation
Methods
We used label-free mass spectrometry to identify differences in signaling pathways after irradiation in normal and radiosensitive individuals. Cell viability was initially determined by water soluble tetrazolium (WST) assay and DNA damage response was analyzed with 53BP1 repair foci formation along with KRAB-associated protein 1 (KAP1) phosphorylation.
Results
Proteomic analysis identified 4028 proteins, which were used in subsequent in silico pathway enrichment analysis to predict affected biological pathways post-IR. In AT cells, networks were heterogeneous at both time points with no common pathway identified. Mitotic cell cycle progress was the most prominent pathway altered after IR in cells from healthy donors. In particular, components of the chromosome passenger complex (INCENP and CDCA8) were significantly downregulated after 72 hours. This could also be verified at the mRNA level.
Conclusion
Altogether, the most striking result was that proteins forming the chromosome passenger complex were downregulated after radiation exposure in healthy normosensitive control cells, but not in radiosensitive ATM-deficient cells. Thus, mitosis-associated proteins form an interesting compound to gain insights into the development and prediction of radiosensitivity.
Objectives
Introducing SPINEPS, a deep learning method for semantic and instance segmentation of 14 spinal structures (ten vertebra substructures, intervertebral discs, spinal cord, spinal canal, and sacrum) in whole-body sagittal T2-weighted turbo spin echo images.
Material and methods
This local ethics committee-approved study utilized a public dataset (train/test 179/39 subjects, 137 female), a German National Cohort (NAKO) subset (train/test 1412/65 subjects, mean age 53, 694 female), and an in-house dataset (test 10 subjects, mean age 70, 5 female). SPINEPS is a semantic segmentation model, followed by a sliding window approach utilizing a second model to create instance masks from the semantic ones. Segmentation evaluation metrics included the Dice score and average symmetrical surface distance (ASSD). Statistical significance was assessed using the Wilcoxon signed-rank test.
Results
On the public dataset, SPINEPS outperformed a nnUNet baseline on every structure and metric (e.g., an average over vertebra instances: dice 0.933 vs 0.911, p < 0.001, ASSD 0.21 vs 0.435, p < 0.001). SPINEPS trained on automated annotations of the NAKO achieves an average global Dice score of 0.918 on the combined NAKO and in-house test split. Adding the training data from the public dataset outperforms this (average instance-wise Dice score over the vertebra substructures 0.803 vs 0.778, average global Dice score 0.931 vs 0.918).
Conclusion
SPINEPS offers segmentation of 14 spinal structures in T2w sagittal images. It provides a semantic mask and an instance mask separating the vertebrae and intervertebral discs. This is the first publicly available algorithm to enable this segmentation.
Key Points
Question No publicly available automatic approach can yield semantic and instance segmentation masks for the whole spine (including posterior elements) in T2-weighted sagittal TSE images .
Findings Segmenting semantically first and then instance-wise outperforms a baseline trained directly on instance segmentation. The developed model produces high-resolution MRI segmentations for the whole spine .
Clinical relevance This study introduces an automatic approach to whole spine segmentation, including posterior elements, in arbitrary fields of view T2w sagittal MR images, enabling easy biomarker extraction, automatic localization of pathologies and degenerative diseases, and quantifying analyses as downstream research .
In regenerating tissues, synthesis and remodeling of membranes rely on lipid turnover and transport. Our study addresses lipid adaptations in intestinal regeneration of Drosophila melanogaster and limb regeneration of Ambystoma mexicanum . We found changes in lipid profiles at different locations: transport, storage organs and regenerating tissues. We demonstrate that attenuating insulin signaling, exclusively in fat storage, inhibits the regeneration-specific response in both the fat storage and the regenerating tissue in Drosophila. Furthermore, in uninjured axolotls we found sex-specific lipid profiles in both storage and circulation, while in regenerating animals these differences subside. The regenerating limb presents a unique sterol profile, albeit with no sex differences. We postulate that regeneration triggers a systemic response, where organs storing lipids play a significant role in the regulation of systemic lipid traffic. Second, that this response may be an active and well-regulated mechanism, as observed when homeostatic sex-differences disappear in regenerating salamanders.
Background
Epigenome‐wide association studies have identified multiple DNA methylation sites (CpGs) associated with alcohol consumption, an important lifestyle risk factor for cardiovascular diseases. This study aimed to test the hypothesis that an alcohol consumption epigenetic risk score (ERS) is associated with blood pressure (BP) traits.
Results
We implemented an ERS based on a previously reported epigenetic signature of 144 alcohol-associated CpGs in meta-analysis of participants of European ancestry. We found a one-unit increment of ERS was associated with eleven drinks of alcohol consumed per day, on average, across several cohorts (p < 0.0001). We examined the association of the ERS with systolic blood pressure (SBP), diastolic blood pressure (DBP), and hypertension (HTN) in 3,898 Framingham Heart Study (FHS) participants. Cross-sectional analyses in FHS revealed that a one-unit increment of the ERS was associated with 1.93 mm Hg higher SBP (p = 4.64E−07), 0.68 mm Hg higher DBP (p = 0.006), and an odds ratio of 1.78 for HTN (p < 2E−16). Meta-analysis of the cross-sectional association of the ERS with BP traits in eight independent external cohorts (n = 11,544) showed similar relationships with BP levels, i.e., a one-unit increase in ERS was associated with 0.74 mm Hg (p = 0.002) higher SBP and 0.50 mm Hg (p = 0.0006) higher DBP, but not with HTN. Longitudinal analyses in FHS (n = 3260) and five independent external cohorts (n = 4021) showed that the baseline ERS was not associated with a change in BP over time or with incident HTN.
Conclusions
Our findings demonstrate that the ERS has potential clinical utility in assessing lifestyle factors related to cardiovascular risk, especially when self-reported behavioral data (e.g., alcohol consumption) are unreliable or unavailable.
Graphic Abstract
Iron sustains cancer cell phenotypic and metabolic plasticity, yet it also sensitizes the mesenchymal/drug-tolerant persister phenotype to ferroptosis. This posits that iron compartmentalization must be tightly regulated. However, the molecular machinery governing organelle Fe2+ compartmentalization remains elusive. Here, we identified Bdh2, the mammalian homolog of the bacterial EntA, as a key effector of inter-organelle Fe2+ redistribution and ferroptosis vulnerability during melanoma transition from a melanocytic (MEL) to a mesenchymal-like (MES) phenotype. In metabolically proficient MEL cells, Bdh2 localizes at the mitochondria-lysosome contacts to generate the siderophore 2,5-dihydroxybenzoic acid (2,5-DHBA) that ferries iron into the mitochondria. Fe2+ transfer by Bdh2 endorses OXPHOS and ATP production, utilized by V-ATPase for lysosomal acidification and MLC maintenance. Loss of Bdh2 expression alters lysosomal pH and MLC tethering dynamics causing lysosomal iron sequestration, which primes MES cells for ferroptosis. Rescuing Bdh2 expression, or supplementing 2,5-DHBA, rectifies lysosomal pH and MLCs, protecting MES cells from ferroptosis and enhancing their ability to metastasize through the bloodstream. Thus, we unveiled a Bdh2-dependent evolutionary-conserved mechanism that orchestrates inter-organelle Fe2+ transfer, linking metabolic regulation of lysosomal pH to the ferroptosis vulnerability of the MES/drug-tolerant persister cells.
Estimating parameters of dynamic models from experimental data is a challenging, and often computationally-demanding task. It requires a large number of model simulations and objective function gradient computations, if gradient-based optimization is used. In many cases, steady-state computation is a part of model simulation, either due to steady-state data or an assumption that the system is at steady state at the initial time point. Various methods are available for steady-state and gradient computation. Yet, the most efficient pair of methods (one for steady states, one for gradients) for a particular model is often not clear. In order to facilitate the selection of methods, we explore six method pairs for computing the steady state and sensitivities at steady state using six real-world problems. The method pairs involve numerical integration or Newton’s method to compute the steady-state, and—for both forward and adjoint sensitivity analysis—numerical integration or a tailored method to compute the sensitivities at steady-state. Our evaluation shows that all method pairs provide accurate steady-state and gradient values, and that the two method pairs that combine numerical integration for the steady-state with a tailored method for the sensitivities at steady-state were the most robust, and amongst the most computationally-efficient. We also observed that while Newton’s method for steady-state computation yields a substantial speedup compared to numerical integration, it may lead to a large number of simulation failures. Overall, our study provides a concise overview across current methods for computing sensitivities at steady state. While our study shows that there is no universally-best method pair, it also provides guidance to modelers in choosing the right methods for a problem at hand.
The contribution of deubiquitylating enzymes (DUBs) to β-Catenin stabilization in intestinal stem cells and colorectal cancer (CRC) is poorly understood. Here, and by using an unbiassed screen, we discovered that the DUB USP10 stabilizes β-Catenin specifically in APC-truncated CRC in vitro and in vivo. Mechanistic studies, including in vitro binding together with computational modelling, revealed that USP10 binding to β-Catenin is mediated via the unstructured N-terminus of USP10 and is outcompeted by intact APC, favouring β-catenin degradation. However, in APC-truncated cancer cells USP10 binds to β-catenin, increasing its stability which is critical for maintaining an undifferentiated tumour identity. Elimination of USP10 reduces the expression of WNT and stem cell signatures and induces the expression of differentiation genes. Remarkably, silencing of USP10 in murine and patient-derived CRC organoids established that it is essential for NOTUM signalling and the APC super competitor-phenotype, reducing tumorigenic properties of APC-truncated CRC. These findings are clinically relevant as patient-derived organoids are highly dependent on USP10, and abundance of USP10 correlates with poorer prognosis of CRC patients. Our findings reveal, therefore, a role for USP10 in CRC cell identity, stemness, and tumorigenic growth by stabilising β-Catenin, leading to aberrant WNT signalling and degradation resistant tumours. Thus, USP10 emerges as a unique therapeutic target in APC truncated CRC.
Dendritic cells (DCs) are crucial for initiating protective immune responses and have also been implicated in the generation and regulation of Foxp3⁺ regulatory T cells (Treg cells). Here, we show that in the lamina propria of the small intestine, the alternative NF-κB family member RelB is necessary for the differentiation of cryptopatch and isolated lymphoid follicle-associated DCs (CIA-DCs). Moreover, single-cell RNA sequencing reveals a RelB-dependent signature in migratory DCs in mesenteric lymph nodes favoring DC-Treg cell interaction including elevated expression and release of the chemokine CCL22 from RelB-deficient conventional DCs (cDCs). In line with the key role of CCL22 to facilitate DC-Treg cell interaction, RelB-deficient DCs have a selective advantage to interact with Treg cells in an antigen-specific manner. In addition, DC-specific RelB knockout animals show increased total Foxp3⁺ Treg cell numbers irrespective of inflammatory status. Consequently, DC-specific RelB knockout animals fail to mount protective Th2-dominated immune responses in the intestine after infection with Heligmosomoides polygyrus bakeri. Thus, RelB expression in cDCs acts as a rheostat to establish a tolerogenic set point that is maintained even during strong type 2 immune conditions and thereby is a key regulator of intestinal homeostasis.
Background
The renal epithelial sodium channel (ENaC) is essential for sodium balance and blood pressure control. ENaC undergoes complex proteolytic activation by not yet clearly identified tubular proteases. Here, we examined a potential role of transmembrane serine protease 2 (TMPRSS2).
Methods
Murine ENaC and TMPRSS2 were (co-)expressed in Xenopus laevis oocytes. ENaC cleavage and function were studied in TMPRSS2-deficient murine cortical collecting duct (mCCD cl1 ) cells and TMPRSS2-knockout (Tmprss2 −/− ) mice. Short-circuit currents (I SC ) were measured to assess ENaC-mediated transepithelial sodium transport of mCCD cl1 cells. The mCCD cl1 cell transcriptome was studied using RNA sequencing. The effect of low-sodium diet with or without high potassium were compared in Tmprss2 −/− and wildtype mice using metabolic cages. ENaC-mediated whole-cell currents were recorded from microdissected tubules of Tmprss2 −/− and wildtype mice.
Results
In oocytes, co-expression of murine TMPRSS2 and ENaC resulted in fully cleaved γ-ENaC and ∼2-fold stimulation of ENaC currents. High baseline expression of TMPRSS2 was detected in mCCD cl1 cells without a stimulatory effect of aldosterone on its function or transcription. TMPRSS2 knockout in mCCD cl1 cells compromised γ-ENaC cleavage and reduced baseline and aldosterone-stimulated I SC which could be rescued by chymotrypsin. A compensatory transcriptional upregulation of other proteases was not observed. Tmprss2 −/− mice kept on standard diet exhibited no apparent phenotype, but renal γ-ENaC cleavage was altered. In response to a low-salt diet, particularly with high potassium intake, Tmprss2 −/− mice increased plasma aldosterone significantly more than wildtype mice to achieve a similar reduction of renal sodium excretion. Importantly, the stimulatory effect of trypsin on renal tubular ENaC currents was much more pronounced in Tmprss2 −/− mice than that in wildtype mice. This indicated the presence of incompletely cleaved and less active channels at the cell surface of TMPRSS2-deficient tubular epithelial cells.
Conclusions
TMPRSS2 contributes to proteolytic ENaC activation in mouse kidney in vivo.
Self-help can play an important supplementary role in the treatment of people with severe mental illness; however, little is known about the utilization of the various approaches.
This study describes the use of various self-help options by patients with severe mental illness and examines potential predictors.
As part of the observational cross-sectional study on patients with severe mental illness (IMPPETUS, N = 397), trained staff collected sociodemographic, illness-associated and treatment-associated data between March 2019 and September 2019. Binary logistic regression was used to analyze a possible association with the use of self-help.
The participants most frequently reported using self-help literature (n = 170; 45.5%) followed by self-help groups (n = 130; 33.2%), electronic mental health applications (n = 56; 15.5%) and self-management approaches (n = 54; 14.8%). Trialogue seminars (n = 36; 9.9%) were the least used by the participants. The utilization of the various approaches is influenced by sociodemographic and disease-related characteristics (age, education, marital status, migration background, age at onset of initial mental health problems, psychosocial functioning level) but not by factors associated with treatment.
The potential of self-help is not being fully utilized in the sample investigated. The reported use of self-help approaches by the participants ranged between 10% and 46%. The various formats address specific target groups. More targeted information must be provided about the various options and the use of self-help in routine treatment must be actively fostered in order to increase the utilization of self-help.
First small sample studies indicate that disturbances of spinal morphology may impair craniospinal flow of cerebrospinal fluid and result in neurodegeneration. The aim of this study was to evaluate the association of cervical spinal canal width and scoliosis with gray matter, white matter, ventricular, and white matter hyperintensity volumes of the brain in a large study sample.
400 participants underwent whole-body 3 tesla magnetic resonance imaging. Gray matter, white matter, and ventricular volumes were quantified using a warp-based automated brain volumetric approach. Cervical spinal canal diameters were measured manually at the 2/3 level. Scoliosis was evaluated using manual measurements of the Cobb angle. Linear binomial regression analyses of measures of brain volumes and spine anatomy were performed while adjusting for age, sex, hypertension, cholesterol levels, body mass index, smoking and alcohol consumption.
383 participants were included (57% male; age: 56.3 (±9.2) years). After adjustment, smaller spinal canal width at the cervical vertebrae 2/3 level was associated with lower gray matter (p=0.034), lower white matter (p=0.012), and higher ventricular (p=0.006, inverse association) volume. Participants with scoliosis had lower gray matter (p=0.005), lower white matter (p=0.011) and larger brain ventricular (p=0.003) volumes than participants without scoliosis. However, these associations were attenuated after adjustment. Spinal canal width at the cervical 2/3 level and scoliosis were not associated with white matter hyperintensity volume before and after adjustment (p>0.864).
In our study cohort smaller spinal canal width at the cervical 2/3 level and scoliosis were associated with lower gray and white matter volumes and larger ventricle size. These characteristics of the spine might constitute independent risk factors for neurodegeneration.
Simultaneous profiling of single-cell gene expression and lineage history holds enormous potential for studying cellular decision-making. Recent computational approaches combine both modalities into cellular trajectories; however, they cannot make use of all available lineage information in destructive time-series experiments. Here, we present moslin, a Gromov-Wasserstein-based model to couple cellular profiles across time points based on lineage and gene expression information. We validate our approach in simulations and demonstrate on Caenorhabditis elegans embryonic development how moslin predicts fate probabilities and putative decision driver genes. Finally, we use moslin to delineate lineage relationships among transiently activated fibroblast states during zebrafish heart regeneration.
- Luis M. García-Marín
- Adrian Campos
- Santiago Diaz Torres
- [...]
- Miguel E. Rentería
Subcortical brain structures are involved in developmental, psychiatric and neurological disorders. Here we performed genome-wide association studies meta-analyses of intracranial and nine subcortical brain volumes (brainstem, caudate nucleus, putamen, hippocampus, globus pallidus, thalamus, nucleus accumbens, amygdala and the ventral diencephalon) in 74,898 participants of European ancestry. We identified 254 independent loci associated with these brain volumes, explaining up to 35% of phenotypic variance. We observed gene expression in specific neural cell types across differentiation time points, including genes involved in intracellular signaling and brain aging-related processes. Polygenic scores for brain volumes showed predictive ability when applied to individuals of diverse ancestries. We observed causal genetic effects of brain volumes with Parkinson’s disease and attention-deficit/hyperactivity disorder. Findings implicate specific gene expression patterns in brain development and genetic variants in comorbid neuropsychiatric disorders, which could point to a brain substrate and region of action for risk genes implicated in brain diseases.
Acute lung injury (ALI) is primarily driven by an intense inflammation in the alveolar epithelium. Key to this is the pro-inflammatory cytokine, Interleukin 17 (IL-17), which influences pulmonary immunity and modifies p53 function. The direct role of IL-17A in p53-fibrinolytic system is still unclear, it is important to evaluate this mechanism to regulate the ALI progression to idiopathic pulmonary fibrosis (IPF). C57BL/6 mice, exposed to recombinant IL-17A protein and treated with curcumin, provided insight into IL-17A mechanisms and curcumin's potential for modulating early pulmonary fibrosis stages. A diverse methodology, including proteomics, single-cell RNA sequencing (scRNA-seq) integration, molecular, and Schroedinger approach were utilized. In silico approaches facilitated the potential interactions between curcumin, IL-17A, and apoptosis-related proteins. A notable surge in the expression levels of IL-17A, p53, and fibrinolytic components such as Plasminogen Activator Inhibitor-1 (PAI-I) was discerned upon the IL17A exposure in mouse lungs. Furthermore, the enrichment of pathways and differential expression of proteins underscored the significance of IL-17A in governing downstream regulatory pathways such as inflammation, NF-kappaB signaling, Mitogen-Activated Protein Kinases (MAPK), p53, oxidative phosphorylation, JAK-STAT, and apoptosis. The integration of scRNA-seq data from 20 IPF and 10 control lung specimens emphasized the importance of IL-17A mediated downstream regulation in PF patients. A potent immuno-pharmacotherapeutic agent, curcumin, demonstrated a substantial capacity to modulate the lung pathology and molecular changes induced by IL-17A in mouse lungs. Human IPF single cell data integration confirmed the effects of IL-17A mediated fibrinolytic components in ALI to IPF progression.
The identification of gene regulatory networks (GRNs) is crucial for understanding cellular differentiation. Single-cell RNA sequencing data encode gene-level covariations at high resolution, yet data sparsity and high dimensionality hamper accurate and scalable GRN reconstruction. To overcome these challenges, we introduce NetID leveraging homogenous metacells while avoiding spurious gene–gene correlations. Benchmarking demonstrates superior performance of NetID compared to imputation-based methods. By incorporating cell fate probability information, NetID facilitates the prediction of lineage-specific GRNs and recovers known network motifs governing bone marrow hematopoiesis, making it a powerful toolkit for deciphering gene regulatory control of cellular differentiation from large-scale single-cell transcriptome data.
Institution pages aggregate content on ResearchGate related to an institution. The members listed on this page have self-identified as being affiliated with this institution. Publications listed on this page were identified by our algorithms as relating to this institution. This page was not created or approved by the institution. If you represent an institution and have questions about these pages or wish to report inaccurate content, you can contact us here.
Information
Address
München, Germany
Head of institution
Prof. Dr. med. Dr. h.c. Matthias H. Tschöp, Heinrich Baßler, Dr. rer. nat. Alfons Enhsen