Cell lineage detection refers to the inference of differentiation pathways from immature cells to distinct mature cell types. We developed TimeFlow 2, a new method for lineage inference in large flow cytometry datasets. It uses a single static snapshot of unordered cells and does not require prior knowledge of the number of pathways, cell types or temporal labels. TimeFlow 2 uses the cell orderings from TimeFlow and defines coarse cell states along pseudotime segments. By connecting these states, it constructs paths at the cell state level. To approximate the trajectory structure, it further groups the paths based on an optimal transport-based cost function. We used TimeFlow 2 on three healthy bone marrow samples and accurately assigned monocytes, neutrophils, erythrocytes and B-cells of different maturation stages to four distinct pathways. Marker dynamics across the inferred pathways showed highly correlated patterns for the corresponding lineages in all three patients. We compared the performance of TimeFlow 2 and three other established methods using standard classification and correlation metrics. TimeFlow 2 outperformed the others on flow cytometry datasets and remained competitive on the challenging mass cytometry datasets. Overall, TimeFlow 2 detects biologically informative pathways, allowing bioinformaticians to model and compare marker dynamics across cell lineages in a data-driven way. Source code in Python and tutorials are available at https://github.com/MargaritaLiarou1/TimeFlow2 .

Background:Risk of bias (RoB) assessment of randomized clinical trials (RCTs) is vital to answering systematic review questions accurately. Manual RoB assessment for hundreds of RCTs is a cognitively demanding and lengthy process. Automation has the potential to assist reviewers in rapidly identifying text descriptions in RCTs that indicate potential risks of bias. However, no RoB text span annotated corpus could be used to fine-tune or evaluate large language models (LLMs), and there are no established guidelines for annotating the RoB spans in RCTs. Objective:The revised Cochrane RoB 2 test (RoB 2) tool provides comprehensive guidelines for RoB assessment; however, due to the inherent subjectivity of this tool, it cannot be directly used as RoB annotation guidelines. The study aimed to develop precise RoB text span annotation instructions that could address this subjectivity and thus aid the corpus annotation. Methods:We leveraged RoB 2 guidelines to develop visual instructional placards that serve as annotation guidelines for RoB spans and risk judgments. Expert annotators used these visual placards to annotate a dataset named RoBuster, consisting of 41 full-text RCTs from the domains of physiotherapy and rehabilitation. We report interannotator agreement (IAA) between 2 annotators for text span annotations before and after applying visual instructions on a subset (n=9) of RoBuster. We also provide IAA on bias risk judgments using Cohen κ. Moreover, we used a portion of RoBuster (n=10) to evaluate an LLM using a straightforward evaluation framework. This evaluation aimed to gauge the performance of an LLM (here GPT 3.5) in the challenging task of RoB span extraction and demonstrate the utility of this corpus using a straightforward framework. Results:We present a corpus of 41 RCTs with fine-grained text span annotations comprising more than 28,427 tokens belonging to 22 RoB classes. The IAA at the text span level calculated using the F1 measure varies from 0% to 90%, while Cohen κ for risk judgments ranges between -0.235 and 1.0. Using visual instructions for annotation increases the IAA by more than 17 percentage points. LLM (GPT-3.5) shows promising but varied observed agreements with the expert annotation across the different bias questions. Conclusions:Despite having comprehensive bias assessment guidelines and visual instructional placards, RoB annotation remains a complex task. Using visual placards for bias assessment and annotation enhances IAA compared to cases where visual placards are absent; however, text annotation remains challenging for the subjective questions and the questions for which annotation data are unavailable in RCTs. Similarly, while GPT-3.5 demonstrates effectiveness, its accuracy diminishes with more subjective RoB questions and low information availability.

Curvature notions on graphs provide a theoretical description of graph topology, highlighting bottlenecks and denser connected regions. Artifacts of the message passing paradigm in Graph Neural Networks, such as oversmoothing and oversquashing, have been attributed to these regions. However, it remains unclear how the topology of a graph interacts with the learned preferences of GNNs. Through Massive Activations, which correspond to extreme edge activation values in Graph Transformers, we probe this correspondence. Our findings on synthetic graphs and molecular benchmarks reveal that MAs do not preferentially concentrate on curvature extremes, despite their theoretical link to information flow. On the Long Range Graph Benchmark, we identify a systemic curvature shift: global attention mechanisms exacerbate topological bottlenecks, drastically increasing the prevalence of negative curvature. Our work reframes curvature as a diagnostic probe for understanding when and why graph learning fails.

Graph Neural Networks (GNNs) have become increasingly popular for effectively modeling graph-structured data, and attention mechanisms have been pivotal in enabling these models to capture complex patterns. In our study, we reveal a critical yet underexplored consequence of integrating attention into edge-featured GNNs: the emergence of Massive Activations (MAs) within attention layers. By developing a novel method for detecting MAs on edge features, we show that these extreme activations are not only activation anomalies but encode domain-relevant signals. Our post-hoc interpretability analysis demonstrates that, in molecular graphs, MAs aggregate predominantly on common bond types (e.g., single and double bonds) while sparing more informative ones (e.g., triple bonds). Furthermore, our ablation studies confirm that MAs can serve as natural attribution indicators, reallocating to less informative edges. Our study assesses various edge-featured attention-based GNN models using benchmark datasets, including ZINC, TOX21, and PROTEINS. Key contributions include (1) establishing the direct link between attention mechanisms and MAs generation in edge-featured GNNs, (2) developing a robust definition and detection method for MAs enabling reliable post-hoc interpretability. Overall, our study reveals the complex interplay between attention mechanisms, edge-featured GNNs model, and MAs emergence, providing crucial insights for relating GNNs internals to domain knowledge.

To assess the performance of automatic rib fracture detection of an existing deep learning (DL) model, nnDetection, on a postmortem (PM) CT scan dataset, and to identify the main factors of domain shift between clinical and PM CT imaging. Rib fracture detection and classification in forensic investigations is a time-consuming yet crucial task that can contribute to determine the cause of death. DL models are a promising tool, as recent research shows that radiologists using DL models can detect rib fractures in clinical CT scans at higher sensitivity and in shorter time. A dataset of 50 PMCT scans (24

Generating high-fidelity and biologically plausible synthetic single-cell RNA sequencing (scRNA-seq) data, especially with conditional control, is challenging due to its high dimensionality, sparsity, and complex biological variations. Existing generative models often struggle to capture these unique characteristics and ensure robustness to structural noise in cellular networks. We introduce LapDDPM, a novel conditional Graph Diffusion Probabilistic Model for robust and high-fidelity scRNA-seq generation. LapDDPM uniquely integrates graph-based representations with a score-based diffusion model, enhanced by a novel spectral adversarial perturbation mechanism on graph edge weights. Our contributions are threefold: we leverage Laplacian Positional Encodings (LPEs) to enrich the latent space with crucial cellular relationship information; we develop a conditional score-based diffusion model for effective learning and generation from complex scRNA-seq distributions; and we employ a unique spectral adversarial training scheme on graph edge weights, boosting robustness against structural variations. Extensive experiments on diverse scRNA-seq datasets demonstrate LapDDPM's superior performance, achieving high fidelity and generating biologically-plausible, cell-type-specific samples. LapDDPM sets a new benchmark for conditional scRNA-seq data generation, offering a robust tool for various downstream biological applications.

Pseudotime methods order cells undergoing differentiation from the least to the most differentiated. We developed TimeFlow, a new method for computing pseudotime in multi-dimensional flow cytometry datasets. TimeFlow tracks the differentiation path of each cell on a graph by following smooth changes in the cell population density. To compute the probability density function of the cells, it uses a normalizing flow model. We profiled bone marrow samples from three healthy patients using a 20-color antibody panel for flow cytometry and prepared datasets that ranged from 5,000 to 600,000 cells and included monocytes, neutrophils, erythrocytes, and B-cells at various maturation stages. TimeFlow computed fine-grained pseudotime for all the datasets, and the cell orderings were consistent with prior knowledge of human hematopoiesis. Experiments showed its potential in generalizing across patients and unseen cell states. We compared our method to 11 other pseudotime methods using in-house and public datasets and found very good performance for both linear and branching trajectories. TimeFlow's pseudotemporal orderings are useful for modeling the dynamics of cell surface proteins along linear trajectories. The biologically meaningful results in branching trajectories suggest the possibility of future applications with automated cell lineage detection. Code is available at https://github.com/MargaritaLiarou1/TimeFlow and data at https://osf.io/ykue7/.

The widespread adoption of Electronic Health Records (EHRs) and deep learning, particularly through Self-Supervised Representation Learning (SSRL) for categorical data, has transformed clinical decision-making. This scoping review, following PRISMA-ScR guidelines, examines 46 studies published from January 2019 to April 2024, sourced from PubMed, MEDLINE, Embase, ACM, and Web of Science, focusing on SSRL for unlabeled categorical EHR data. The review systematically assesses research trends in building computationally and data-efficient representations for medical tasks, identifying major trends in model families: Transformer-based (43%), Autoencoder-based (28%), and Graph Neural Network-based (17%) models. The analysis highlights scenarios where healthcare institutions can leverage or develop SSRL technologies. It also addresses current limitations in assessing the impact of these technologies and identifies research opportunities to enhance their influence on clinical practice.

Dimensionality reduction (DR) methods are often used to unveil important information from a dataset. Most advanced DR algorithms are non-linear, presenting better performance in the result at the cost of a lower interpretability of whether the reduced dataset resembles the original one. While new methods for dimensionality reduction arise, UMAP is still more used than any other, for its great performance and solid theoretical foundations. As an embedding operation, DR methods induce a loss of information. Evaluation measures can be used in order to quantify the (local) faithfulness of the reduction. In this paper we propose using fuzzy simplicial sets (the cornerstone of UMAP) to develop evaluation measures. We demonstrate the usefulness of these measures in detecting unfaithfulness by revealing structural distortions, and identifying when the reduction falls below the intrinsic dimensionality.

We present LaplaceGNN, a novel self-supervised graph learning framework that bypasses the need for negative sampling by leveraging spectral bootstrapping techniques. Our method integrates Laplacian-based signals into the learning process, allowing the model to effectively capture rich structural representations without relying on contrastive objectives or handcrafted augmentations. By focusing on positive alignment, LaplaceGNN achieves linear scaling while offering a simpler, more efficient, self-supervised alternative for graph neural networks, applicable across diverse domains. Our contributions are twofold: we precompute spectral augmentations through max-min centrality-guided optimization, enabling rich structural supervision without relying on handcrafted augmentations, then we integrate an adversarial bootstrapped training scheme that further strengthens feature learning and robustness. Our extensive experiments on different benchmark datasets show that LaplaceGNN achieves superior performance compared to state-of-the-art self-supervised graph methods, offering a promising direction for efficiently learning expressive graph representations.

The computation of a continuous generative model to describe a finite sample of an infinite metric space can prove challenging and lead to erroneous hypothesis, particularly in high-dimensional spaces. In this paper, we follow a different route and define the Hubness Half Space Partitioning graph (HubHSP graph). By constructing this spanning graph over the dataset, we can capture both the geometrical and statistical properties of the data without resorting to any continuity assumption. Leveraging the classical graph-theoretic apparatus, the HubHSP graph facilitates critical operations, including the creation of a representative sample of the original dataset, without relying on density estimation. This representative subsample is essential for a range of operations, including indexing, visualization, and machine learning tasks such as clustering or inductive learning. With the HubHSP graph, we can bypass the limitations of traditional methods and obtain a holistic understanding of our dataset’s properties, enabling us to unlock its full potential.

In the realm of hematologic cell populations classification, the intricate patterns within flow cytometry data necessitate advanced analytical tools. This paper presents 'HemaGraph', a novel framework based on Graph Attention Networks (GATs) for single-cell multi-class classification of hematological cells from flow cytometry data. Harnessing the power of GATs, our method captures subtle cell relationships, offering highly accurate patient profiling. Based on evaluation of data from 30 patients, HemaGraph demonstrates classification performance across five different cell classes, outperforming traditional methodologies and state-of-the-art methods. Moreover, the uniqueness of this framework lies in the training and testing phase of HemaGraph, where it has been applied for extremely large graphs, containing up to hundreds of thousands of nodes and two million edges, to detect low frequency cell populations (e.g. 0.01 underscore the potential of HemaGraph in improving hematoligic multi-class classification, paving the way for patient-personalized interventions. To the best of our knowledge, this is the first effort to use GATs, and Graph Neural Networks (GNNs) in general, to classify cell populations from single-cell flow cytometry data. We envision applying this method to single-cell data from larger cohort of patients and on other hematologic diseases.

In the complex landscape of hematologic samples such as peripheral blood or bone marrow derived from flow cytometry (FC) data, cell-level prediction presents profound challenges. This work explores injecting hierarchical prior knowledge into graph neural networks (GNNs) for single-cell multi-class classification of tabular cellular data. By representing the data as graphs and encoding hierarchical relationships between classes, we propose our hierarchical plug-in method to be applied to several GNN models, namely, FCHC-GNN, and effectively designed to capture neighborhood information crucial for single-cell FC domain. Extensive experiments on our cohort of 19 distinct patients, demonstrate that incorporating hierarchical biological constraints boosts performance significantly across multiple metrics compared to baseline GNNs without such priors. The proposed approach highlights the importance of structured inductive biases for gaining improved generalization in complex biological prediction tasks.

This paper presents FlowCyt, the first comprehensive benchmark for multi-class single-cell classification in flow cytometry data. The dataset comprises bone marrow samples from 30 patients, with each cell characterized by twelve markers. Ground truth labels identify five hematological cell types: T lymphocytes, B lymphocytes, Monocytes, Mast cells, and Hematopoietic Stem/Progenitor Cells (HSPCs). Experiments utilize supervised inductive learning and semi-supervised transductive learning on up to 1 million cells per patient. Baseline methods include Gaussian Mixture Models, XGBoost, Random Forests, Deep Neural Networks, and Graph Neural Networks (GNNs). GNNs demonstrate superior performance by exploiting spatial relationships in graph-encoded data. The benchmark allows standardized evaluation of clinically relevant classification tasks, along with exploratory analyses to gain insights into hematological cell phenotypes. This represents the first public flow cytometry benchmark with a richly annotated, heterogeneous dataset. It will empower the development and rigorous assessment of novel methodologies for single-cell analysis.

Data arising from sensors is generally high-dimensional and manifold learning or, more generally, embedding techniques are applied for dimension reduction, possibly with the hope of circumventing the curse of dimensionality. Low-dimensional data may then be the basis for further processing, such as clustering or learning. It is therefore critical that the reduced data representation is faithful to the information contained in the original data. Manifold learning methods are generally evaluated either by visual inspection or by quantifying globally the preservation of neighborhood structures over known dataset. In this paper, we argue for measures that behave smoothly along increasing unfaithfulness to the original data. To build such measures, we return to the manifold assumption and exploit topological information. We further and principally argue for the utility of local measurements of unfaithfulness of representation, as distortions may not be distributed uniformly over the data. The aim here is less to compare manifold techniques than to assess a given technique for its faithfulness. Experiments demonstrate the value of our proposals.

Objectives To review studies on deep learning (DL) models for classification, detection, and segmentation of rib fractures in CT data, to determine their risk of bias (ROB), and to analyse the performance of acute rib fracture detection models.Methods Research articles written in English were retrieved from PubMed, Embase, and Web of Science in April 2023. A study was only included if a DL model was used to classify, detect, or segment rib fractures, and only if the model was trained with CT data from humans. For the ROB assessment, the Quality Assessment of Diagnostic Accuracy Studies tool was used. The performance of acute rib fracture detection models was meta-analysed with forest plots.Results A total of 27 studies were selected. About 75% of the studies have ROB by not reporting the patient selection criteria, including control patients or using 5-mm slice thickness CT scans. The sensitivity, precision, and F1-score of the subgroup of low ROB studies were 89.60% (95%CI, 86.31%-92.90%), 84.89% (95%CI, 81.59%-88.18%), and 86.66% (95%CI, 84.62%-88.71%), respectively. The ROB subgroup differences test for the F1-score led to a p-value below 0.1.Conclusion ROB in studies mostly stems from an inappropriate patient and data selection. The studies with low ROB have better F1-score in acute rib fracture detection using DL models.Advances in knowledge This systematic review will be a reference to the taxonomy of the current status of rib fracture detection with DL models, and upcoming studies will benefit from our data extraction, our ROB assessment, and our meta-analysis.

In the complex landscape of hematologic samples such as peripheral blood or bone marrow, cell classification, delineating diverse populations into a hierarchical structure, presents profound challenges. This study presents LeukoGraph, a recently developed framework designed explicitly for this purpose employing graph attention networks (GATs) to navigate hierarchical classification (HC) complexities. Notably, LeukoGraph stands as a pioneering effort, marking the application of graph neural networks (GNNs) for hierarchical inference on graphs, accommodating up to one million nodes and millions of edges, all derived from flow cytometry data. LeukoGraph intricately addresses a classification paradigm where for example four different cell populations undergo flat categorization, while a fifth diverges into two distinct child branches, exemplifying the nuanced hierarchical structure inherent in complex datasets. The technique is more general than this example. A hallmark achievement of LeukoGraph is its F-score of 98 outclassing prevailing state-of-the-art methodologies. Crucially, LeukoGraph's prowess extends beyond theoretical innovation, showcasing remarkable precision in predicting both flat and hierarchical cell types across flow cytometry datasets from 30 distinct patients. This precision is further underscored by LeukoGraph's ability to maintain a correct label ratio, despite the inherent challenges posed by hierarchical classifications.

Computational pathology targets the automatic analysis of Whole Slide Images (WSI). WSIs are high-resolution digitized histopathology images, stained with chemical reagents to highlight specific tissue structures and scanned via whole slide scanners. The application of different parameters during WSI acquisition may lead to stain color heterogeneity, especially considering samples collected from several medical centers. Dealing with stain color heterogeneity often limits the robustness of methods developed to analyze WSIs, in particular Convolutional Neural Networks (CNN), the state-of-the-art algorithm for most computational pathology tasks. Stain color heterogeneity is still an unsolved problem, although several methods have been developed to alleviate it, such as Hue-Saturation-Contrast (HSC) color augmentation and stain augmentation methods. The goal of this paper is to present Data-Driven Color Augmentation (DDCA), a method to improve the efficiency of color augmentation methods by increasing the reliability of the samples used for training computational pathology models. During CNN training, a database including over 2 million H&E color variations collected from private and public datasets is used as a reference to discard augmented data with color distributions that do not correspond to realistic data. DDCA is applied to HSC color augmentation, stain augmentation and H&E-adversarial networks in colon and prostate cancer classification tasks. DDCA is then compared with 11 state-of-the-art baseline methods to handle color heterogeneity, showing that it can substantially improve classification performance on unseen data including heterogeneous color variations.

In this paper, we delve into the Mutual k-Nearest Neighbor Graph ( m k NNG ) and its significance in clustering and outlier detection. We present a rigorous mathematical framework elucidating its application and highlight its role in the success of various clustering algorithms. Building on Brito et al.’s findings, which link the connected components of the m k NNG to clusters under specific density bounds, we explore its relevance in the context of a wide range of density functions.

Deep hashing has shown to be a complexity-efficient solution for the Approximate Nearest Neighbor search problem in high dimensional space. Many methods usually build the loss function from pairwise or triplet data points to capture the local similarity structure. Other existing methods construct the similarity graph and consider all points simultaneously. Auto-encoding Twin-bottleneck Hashing is one such method that dynamically builds the graph. Specifically, each input data is encoded into a binary code and a continuous variable, or the so-called twin bottlenecks. The similarity graph is then computed from these binary codes, which get updated consistently during the training. In this work, we generalize the original model into a supervised deep hashing network by incorporating the label information. In addition, we examine the differences of codes structure between these two networks and consider the class imbalance problem especially in multi-labeled datasets. Experiments on three datasets yield statistically significant improvement against the original model. Results are also comparable and competitive to other supervised methods.