Bacterial colonies, as dynamic ecosystems, display intricate behaviors and organizational structures that profoundly influence their survival and functionality. These communities engage in physiological and social interactions, resulting in remarkable spatial heterogeneity. Recent advancements in technology and modeling have significantly enhanced our comprehension of these phenomena, shedding light on the underlying mechanisms governing bacterial colony development. In this review, we explore the multifaceted aspects of bacterial colonies, emphasizing their physiological intricacies, innovative research tools, and predictive modeling approaches. By integrating diverse perspectives, we aim to deepen our understanding of these microbial communities and pave the way for novel applications in biotechnology, ecology, and medicine.

Estimating the transmission fitness of SARS-CoV-2 variants and understanding their evolutionary fitness trends are important for epidemiological forecasting. Existing methods are often constrained by their parametric natures and do not satisfactorily align with the observations during COVID-19. Here, we introduce a sliding-window data-driven pairwise comparison method, the differential population growth rate (DPGR) that uses viral strains as internal controls to mitigate sampling biases. DPGR is applicable in time windows in which the logarithmic ratio of two variant subpopulations is approximately linear. We apply DPGR to genomic surveillance data and focus on variants of concern (VOCs) in multiple countries and regions. We found that the log-linear assumption of DPGR can be reliably found within appropriate time windows in many areas. We show that DPGR estimates of VOCs align well with regional empirical observations in different countries. We show that DPGR estimates agree with another method for estimating pathogenic transmission. Furthermore, DPGR allowed us to construct viral relative fitness landscapes that capture the shifting trends of SARS-CoV-2 evolution, reflecting the relative changes of transmission traits for key genotypic changes represented by major variants. The straightforward log-linear regression approach of DPGR may also facilitate its easy adoption. This study shows that DPGR is a promising new tool in our repertoire for addressing future pandemics.

Cellular plasticity enables cells to dynamically adapt to environmental changes by altering their phenotype. This plasticity plays a crucial role in tissue repair and regeneration and contributes to pathological processes such as cancer metastasis. Advances in single-cell omics have significantly advanced the study of cellular states and provided new opportunities for accurate cell classification and uncovering cellular transitions. In this perspective, we emphasize integrating chromatin accessibility data and extrinsic factors, such as microenvironmental cues, with single-cell transcriptomic data to develop holistic models for identifying plastic cell states. Additionally, coupling artificial intelligence with single-cell omics offers transformative potential to address existing challenges and fill gaps in identifying and characterizing plastic cells. We envision the development of a universal plasticity metric, a standardized metric for quantifying cellular plasticity. This metric would enable consistent measurement across diverse studies, creating a unified framework that bridges fields such as developmental biology, cancer research, and regenerative medicine. Fostering innovative approaches to identifying and analyzing cellular plasticity promises not only to deepen our understanding of cellular plasticity but also to accelerate therapeutic advancements, paving the way for novel precision medicine strategies to treat complex diseases such as cancer.

Non-coding RNAs (ncRNAs) have emerged as key regulators in tumorigenesis. In this perspective, we briefly review the significance of ncRNA in cancer biology and highlight recent technological advancements in characterization of ncRNA in cancer research. Specifically, we discuss how these advanced approaches, such as Patho-DBiT, CRISPR screens, and snoKARR-seq, hold the potential to revolutionize ncRNA research by offering comprehensive insights into their spatial expression patterns and functional roles.

Cell lineage tracing is a crucial technique for understanding cell fate and lineage relationships, with wide applications in developmental biology, tissue regeneration, and disease progression studies. Over the years, experimental cell lineage tracing methods have advanced from early labeling techniques to modern genetic tools such as CRISPR-Cas9-based barcoding, whereas computational methods have emerged to analyze high-dimensional data from single-cell sequencing and other omics technologies. This paper reviews both experimental and computational methods, highlighting their respective strengths, limitations, and synergies. Experimental techniques focus on labeling and tracking cells, whereas computational approaches reconstruct lineage relationships and model cellular dynamics. Despite significant progress, challenges remain, including issues with accuracy, resolution, multi-omics integration, and scalability. Future directions involve improvements in experimental techniques and the development of computational methods enhanced by advancements in artificial intelligence. These innovations are expected to drive the field forward, offering potential applications in uncovering the mysteries of life.

Cellular aging is a multifaceted complex process. Many genes and factors have been identified that regulate cellular aging. However, how these genes and factors interact with one another and how these interactions drive the aging processes in single cells remain largely unclear. Recently, computational systems biology has demonstrated its potential to empower aging research by providing quantitative descriptions and explanations of complex aging phenotypes, mechanistic insights into the emergent dynamic properties of regulatory networks, and testable predictions that can guide the design of new experiments and interventional strategies. In general, current complex systems approaches can be categorized into two types: (1) network maps that depict the topologies of large-scale molecular networks without detailed characterization of the dynamics of individual components and (2) dynamical models that describe the temporal behavior in a particular set of interacting factors. In this review, we discuss examples that showcase the application of these approaches to cellular aging with a specific focus on the progress in quantifying and modeling the replicative aging of budding yeast Saccharomyces cerevisiae. We further propose potential strategies for integrating network maps and dynamical models toward a more comprehensive, mechanistic, and predictive understanding of cellular aging. Finally, we outline directions and questions in aging research where systems-level approaches may be especially powerful.

Comorbidity, the co-occurrence of multiple medical conditions in a single patient, profoundly impacts disease management and outcomes. Understanding these complex interconnections is crucial, especially in contexts where comorbidities exacerbate outcomes. Leveraging insights from the human interactome and advancements in graph-based methodologies, this study introduces transformer with subgraph positional encoding (TSPE) for disease comorbidity prediction. Inspired by biologically supervised embedding, TSPE employs transformer's attention mechanisms and subgraph positional encoding (SPE) to capture interactions between nodes and disease associations. Our proposed SPE proves more effective than Laplacian positional encoding, as used in Dwivedi et al.'s graph transformer, underscoring the importance of integrating clustering and disease-specific information for improved predictive accuracy. Evaluated on real clinical benchmark datasets (RR0 and RR1), TSPE demonstrates substantial performance enhancements over the state-of-the-art method, achieving up to 28.24% higher ROC AUC (receiver operating characteristic-area under the curve) and 4.93% higher accuracy. This method shows promise for adaptation to other complex graph-based tasks and applications. The source code is available at GitHub website (xihan-qin/TSPE-GraphTransformer).

In this study, we show an example of a numerical model based on the Keller-Segel system of equations to simulate angiogenesis in response to chemotaxis under Robin boundary conditions, which represent the presence of flux at the tumor. Different parameters of the model are modified to identify key biological factors relevant to the behavior of angiogenesis. The results show that in the presence of a stronger flux, angiogenesis occurs later owing to the chemical flux that creates a more uniform and homogeneous matrix, decreasing the pronunciation of the gradient and reducing the potential of chemotaxis.

Do we need a foundation model (FM) for spatial transcriptomic analysis? To answer this question, we prepared this perspective as a primer. We first review the current progress of developing FMs for modeling spatial transcriptomic data and then discuss possible tasks that can be addressed by FMs. Finally, we explore future directions of developing such models for understanding spatial transcriptomics by describing both opportunities and challenges. In particular, we expect that a successful FM should boost research productivity, increase novel biological discoveries, and provide user-friendly access.

Recent advancements in spatial transcriptomics (ST) technologies allow researchers to simultaneously measure RNA expression levels for hundreds to thousands of genes while preserving spatial information within tissues, providing critical insights into spatial gene expression patterns, tissue organization, and gene functionality. However, existing methods for clustering spatially variable genes (SVGs) into co-expression modules often fail to detect rare or unique spatial expression patterns. To address this, we present spatial transcriptomics iterative hierarchical clustering (stIHC), a novel method for clustering SVGs into co-expression modules, representing groups of genes with shared spatial expression patterns. Through three simulations and applications to ST datasets from technologies such as 10x Visium, 10x Xenium, and Spatial Transcriptomics, stIHC outperforms clustering approaches used by popular SVG detection methods, including SPARK, SPARK-X, MERINGUE, and SpatialDE. Gene ontology enrichment analysis confirms that genes within each module share consistent biological functions, supporting the functional relevance of spatial co-expression. Robust across technologies with varying gene numbers and spatial resolution, stIHC provides a powerful tool for decoding the spatial organization of gene expression and the functional structure of complex tissues.

Cancer is a complex and heterogeneous disease characterized by various genetic and epigenetic alterations. Early diagnosis, accurate subtyping, and staging are essential for effective, personalized treatment and improved survival rates. Traditional diagnostic methods, such as biopsies, are invasive and carry operational risks that hinder repeated use, underscoring the need for noninvasive and personalized alternatives. In response, this study integrates transcriptomic data into human genome-scale metabolic models (GSMMs) to derive patient-specific flux distributions, which are then combined with genomic, proteomic, and fluxomic (JX) data to develop a robust multi-omic classifier for lung cancer subtyping and early diagnosis. The JX classifier is further enhanced by analyzing heterogeneous datasets from RNA sequencing and microarray analyses derived from both tissue samples and cell culture experiments, thereby enabling the identification of key marker features and enriched pathways such as lipid metabolism and energy production. This integrated approach not only demonstrates high performance in distinguishing lung cancer subtypes and early-stage disease but also proves robust when applied to limited pancreatic cancer data. By linking genotype to phenotype, GSMM-driven flux analysis overcomes challenges related to metabolome data scarcity and platform variability by proposing marker processes and reactions for further investigation, ultimately facilitating noninvasive diagnostics and the identification of actionable biomarkers for targeted therapeutic intervention. These findings offer significant promise for streamlining clinical workflows and enabling personalized therapeutic strategies, and they highlight the potential of our versatile workflow for unveiling novel biomarker landscapes in less studied diseases.