Biology subjects

Linsley, D.

Publications and source records attributed to Linsley, D..

3 recordsLinked to original sources

bioRxiv2026

Predictive Cellular Signatures from Live Human Motor Neurons Distinguish TDP-43 ALS and Enable ALS Subtype Stratification

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disorder characterized by the progressive, rapid deterioration of motor neurons (MNs). Rare mutations in a handful of genes are sufficient to cause ALS; however, 90% of ALS cases are not linked to these genes and their underlying cause remains unknown. Abnormal subcellular distribution, structure or aggregation of the TDP-43 protein are nearly universal hallmarks of the disease, suggesting a shared molecular mechanism across both genetic and sporadic ALS (sALS). However, the heterogeneity of the ALS clinical syndrome suggests that the underlying mechanisms culminating in ALS and TDP-43 pathology may partly differ among individuals and may need to be understood to develop successful therapies that target subgroups of patients. Here, we harnessed the power of machine learning (ML) to begin to decode, in a systematic and unbiased fashion, the cellular signatures of ALS. We used high-content imaging of live, human iPSC-derived motor neurons (iMNs) from ALS patients or gene-edited and gene-corrected TDP-43 mutant lines to train shallow connected ML algorithms (SMLs) and deep convolutional neural networks (DNNs). Our models identified and distinguished mutant and control iMNs with moderately high accuracy. We then used explainability methods to uncover the discriminating cellular signals and found that the strongest ones mapped to the nuclear area, suggesting underlying alterations within the nucleus. We validated this finding by revealing that TDP-43 mutant iMNs display alterations in nucleocytoplasmic shuttling and cellular integrity. Further, a time-interaction ML model uncovered dynamic morphological transitions preceding degeneration, offering a window into early pathogenic events as well as neurodevelopmental changes. Extending our ML pipeline to iMNs with mutations in the ALS gene C9orf72 or derived from sALS revealed both overlapping and distinguishable signatures, suggesting shared yet distinct mechanistic pathways. Together, these findings establish ML-driven phenotypic profiling as a powerful approach to stratify people with ALS, help disentangle the molecular heterogeneity of ALS and produce a more holistic phenotypic definition in cell-based models, and ultimately find causes and treatments. This strategy offers a scalable and innovative paradigm for uncovering early disease mechanisms not only in ALS but potentially across a spectrum of neurodegenerative and sporadic disorders.

neuroscience↗

bioRxiv2022

Deceptive learning in histopathology

Deep learning holds immense potential for histopathology, automating tasks that are simple for expert pathologists, and revealing novel biology for tasks that were previously considered difficult or impossible to solve by eye alone. However, the extent to which the visual strategies learned by deep learning models in histopathological analysis are trustworthy or not has yet to be systematically analyzed. In this work, we address this problem and discover new limits on the histopathological tasks for which deep learning models learn trustworthy versus deceptive solutions. While tasks that have been extensively studied in the field like tumor detection are reliable and trustworthy, recent advances demonstrating the ability to learn molecular profiling from hematoxylin and eosin (H&E) stained slides do not hold up to closer scrutiny. Our analysis framework represents a new approach in understanding the capabilities of deep learning models, which should be incorporated into the computational pathologists toolkit.

pathology↗

bioRxiv2020

Super-human cell death detection with biomarker-optimized neural networks

Cell death is an essential process in biology that must be accounted for in live microscopy experiments. Nevertheless, cell death is difficult to detect without perturbing experiments with stains, dyes or biosensors that can bias experimental outcomes, lead to inconsistent results, and reduce the number of processes that can be simultaneously labelled. These additional steps also make live microscopy difficult to scale for high-throughput screening because of the cost, labor, and analysis they entail. We address this fundamental limitation of live microscopy with biomarker-optimized convolutional neural networks (BO-CNN): computer vision models trained with a ground truth biosensor that detect live cells with superhuman, 96% accuracy more than 100 times faster than previous methods. Our models learn to identify important morphological characteristics associated with cell vitality without human input or additional perturbations, and to generalize to other imaging modalities and cell types for which they have no specialized training. We demonstrate that we can interpret decisions from BO-CNN models to gain biological insight into the patterns they use to achieve superhuman accuracy. The BO-CNN approach is broadly useful for live microscopy, and affords a powerful new paradigm for advancing the state of high-throughput imaging in a variety of contexts.

cell biology↗