Biology subjects

Kurjan, A.

Publications and source records attributed to Kurjan, A..

7 recordsLinked to original sources

bioRxiv2026

FlowBench: separating planning, fault recovery and interpretation in agentic bioinformatics

AO_SCPLOWBSTRACTC_SCPLOWAgentic large language model (LLM) systems are being deployed in bioinformatics faster than they are understood, and single-metric evaluations conflate capabilities that fail independently. We introduce FlowBench, a benchmark that decomposes agentic bioinformatics performance into planning, fault recovery, biological interpretation, and end-to-end output-fidelity. Existing systems achieve high plan completeness, but their closed, single-provider designs prevent attribution of performance to scaffolding versus the underlying model. We therefore built FlowAgent, a modular, provider-agnostic framework whose components can be selectively disabled and whose backbone model can be swapped across providers on a shared harness, and used it to evaluate 23 models from three main providers. Three findings emerge. First, generating a valid workflow plan from a named toolchain is largely solved, whereas inferring an appropriate toolchain from biological intent alone is uniformly difficult regardless of model tier, compressing all models into a narrow 44-57% pass-rate band. Second, ablation shows that the dependency-structured plan and a completeness-reflection step drive performance, while adding a same-context validator-driven retry makes structural quality worse. Third, fault recovery and data-grounded interpretation remain unsolved. Models frequently propose fixes that force a clean exit while leaving the underlying data invalid, and data-grounded interpretation lags internal-knowledge recall by a consistent margin. Safety does not emerge from capability, and reasoning-tier models were among the least reliable at recognising unrecoverable faults. Once planning saturates, agent architecture and refusal calibration, not model scale, are the productive frontier. Availability and implementationFlowAgent and FlowBench are available under a GPLv3 licence at https://github.com/EnteloBio/flowagent Contactadam@entelo.bio

bioinformatics↗

bioRxiv2025

Fibroblast specialisation across microanatomy in a single-cell atlas of healthy human Achilles tendon

Tendons are transitional tissues linking muscle to bone, enabling locomotion and fine motor control. The cellular biology across the Achilles tendon unit is poorly understood, yet critical for interpreting normal function and pathological changes across its microanatomically-defined functional zones. We generated a spatially-resolved transcriptomic atlas of human Achilles tendon, sampling the tendon-bone junction (enthesis), midbody, myotendinous junction, and adjoining muscle. Six fibroblast subtypes were identified, with distinct transcriptional profiles and spatial distributions, suggesting specialised functional roles across the tendon-muscle unit. Two dominant fibroblast types were specifically positioned in the tendon mid-substance and paratenon (vessel-rich region surrounding the tendon fibrils); other populations included perineural, myotendinous junction-specific, muscle-specific, and lining-layer fibroblasts. These findings demonstrate how cellular diversity across a transitional tissue may underlie microanatomical-specific roles. This atlas provides a foundation for understanding cellular functions across the tendon-muscle unit and will be essential for comparisons with diseased tissue, identifying pathogenic mediators and treatment targets for autoimmune and degenerative pathologies of the Achilles tendon.

genomics↗

bioRxiv2025

Cellular and molecular landscapes of human tendons across the lifespan revealed by spatial and single-cell transcriptomics

Tendon injuries are common and often heal poorly. While developing tendons heal without scarring, this capacity declines with age, yet the underlying cellular transitions remain poorly defined. Here, we integrate histological, single-nucleus, single-cell, and spatial transcriptomic profiling of human Achilles and quadriceps tendons across embryonic, foetal, and adult stages, including ruptured adult tendons. We identify seven embryonic progenitor states that give rise to three distinct tendon-associated fibrillar, connective tissue, and chondrogenic lineages. These populations diversify during development and occupy distinct spatial niches, adopting specialised roles in matrix synthesis, tissue remodelling, and mechanical adaptation. While non-fibroblast populations remain transcriptionally stable with age, fibroblasts undergo marked reprogramming, shifting to homeostatic or injury-responsive states. In ruptured adult tendons, a subset of fibroblasts partially reactivates developmental programs but remains transcriptionally distinct from their regenerative counterparts. These findings define the cellular architecture of human tendon development and ageing and reveal lineage-specific targets for therapeutic repair.

cell biology↗

bioRxiv2025

FlowAgent: A Modular Agent-Based System for Automated Workflow Management and Data Interpretation

Reproducibility, automation, and flexibility remain persistent challenges in bioinformatics, where complex workflows require integration of diverse computational tools, rigorous quality control, and dynamic adaptability to evolving datasets. Existing workflow managers often trade off usability for flexibility, limiting accessibility for non-specialists and enforcing rigid execution pipelines. We present FlowAgent, an adaptive agent-based framework that embeds computational heuristics and intelligent decision-making into workflow management. Unlike static pipeline managers, FlowAgent leverages agentic systems for context-aware automation, enhancing error detection, correction, and real-time optimisation while minimising manual intervention. By integrating bioinformatics tools with adaptive control, FlowAgent ensures robust quality control and enables workflow execution through an intuitive natural language interface. This approach democratises advanced computational methods for biologists while allowing bioinformaticians to focus on high-level analysis. FlowAgent redefines workflow automation, making bioinformatics pipelines more intelligent, efficient, and accessible, ultimately accelerating scientific discovery across diverse omics data modalities. Availability and implementationFlowAgent is available with GLPv3 license at https://github.com/EnteloBio/flowagent

genomics↗

bioRxiv2024

Exploring cellular changes in ruptured human quadriceps tendons at single-cell resolution

Tendon ruptures in humans have regularly been studied during the chronic phase of injury. However, the early response to injury remains less investigated. Quadriceps tendons, which require prompt surgical treatment, offer a model to investigate this early response. Therefore, this study aimed to explore the early cellular changes in ruptured compared to healthy human quadriceps tendons. Quadriceps tendon samples were collected from patients undergoing tibial shaft fracture repair (healthy) or tendon repair surgery for complete rupture (collected 7-8 days post-injury). Nuclei were isolated for single-nucleus RNA sequencing, and comprehensive transcriptomic analysis was conducted to profile cellular changes. The transcriptomes of 12,808 nuclei were profiled, including 7,268 nuclei from healthy and 5,540 nuclei from ruptured quadriceps tendons, revealing 12 major cell types and several cell subtypes and states. Rupture samples showed increased expression of genes related to extracellular matrix organisation and cell cycle signalling, and a decrease in expression of genes in lipid metabolism pathways. These changes were driven predominantly by gene expression changes in the fibroblast, vascular endothelial cells (VECs), mural cell, and macrophage populations: fibroblasts shift to an activated phenotype upon rupture and there is an increase in proportion of capillary and dividing VECs, suggesting an angiogenic response. A diverse immune environment was observed, with a shift from homeostatic to activated macrophages following rupture. Cell-cell interactions increased in rupture, both in their number and diversity, and primarily involving fibroblast and endothelial cell populations. Collectively, this transcriptomic analysis suggests that fibroblasts and endothelial cells are key orchestrators of the early injury response within ruptured quadriceps tendon.

genomics↗

bioRxiv2022

Single nucleus and spatial transcriptomic profiling of human healthy hamstring tendon

The molecular and cellular basis of health in human tendons remains poorly understood. Amongst human tendons, the hamstrings are the least likely to be injured or degenerate, providing a prototypic healthy tendon reference. The aim of this study was to define the transcriptome and location of all cell types in healthy hamstring tendon. We profiled the transcriptomes of 10,533 nuclei from 4 healthy donors using single-nucleus RNA sequencing (snRNA-seq) and identified 12 distinct cell types. We confirmed the presence of two fibroblast cell types, endothelial cells, mural cells, and immune cells, and revealed the presence of cell types previously unreported for tendon sites, including different skeletal muscle cell types, satellite cells, adipocytes, and nerve cells, which are undefined nervous system cells. Location of these cell types within tendon was defined using spatial transcriptomics and imaging, and transcriptional networks and cell-cell interactions were identified. We demonstrate that fibroblasts have a high number of potential cell-cell interactions, are present throughout the whole tendon tissue, and play an important role in the production and organisation of extracellular matrix, thus confirming their role as key regulators of hamstring tendon tissue homeostasis. Overall, our findings highlight the highly complex cellular networks underpinning tendon function and underpins the importance of fibroblasts as key regulators of hamstring tendon tissue homeostasis.

cell biology↗

bioRxiv2019

The antimicrobial peptide Defensin cooperates with Tumour Necrosis Factor to drive tumour cell death in Drosophila

Antimicrobial peptides (AMPs) are small cationic molecules best known as mediators of the innate defence against microbial infection. While in vitro and ex vivo evidence suggest AMPs capacity to kill cancer cells, in vivo demonstration of an anti-tumour role of endogenous AMPs is lacking. Using a Drosophila model of tumourigenesis, we demonstrate a role for the AMP Defensin in the control of tumour progression. Our results reveal that Tumour Necrosis Factor mediates exposure of phosphatidylserine (PS), which makes tumour cells selectively sensitive to the action of Defensin remotely secreted from tracheal and fat tissues. Defensin binds tumour cells in PS-enriched areas, provoking cell death and tumour regression. Altogether, our results provide the first in vivo demonstration for a role of an endogenous AMP as an anti-cancer agent, as well as a mechanism that explains tumour cell sensitivity to the action of AMPs.

cancer biology↗