Biology subjects

Seifi, M.

Publications and source records attributed to Seifi, M..

5 recordsLinked to original sources

bioRxiv2026

AI4Life Open Calls and Public Challenges: why, how, and what we have learned.

Within AI4Life, we ran three Open Calls and three Public Challenges (2023-2025), supporting 22 bioimage analysis projects from 151 applications and engaging 225 challenge participants, with the aim of applying FAIR deep learning in the life sciences. Our experience offers a view of the current state of bioimage analysis, the landscape of available tools, as well as the existing gaps between method developers, tool producers and potential users. It highlights that even after careful selection for AI-ready projects, most still require substantial effort to apply deep learning, and that the field still relies heavily on established, well-rounded methods to solve common problems. We come to the conclusion that for scientific AI in biology, the rate-limiting step is not methods and models but data, annotations, and shared infrastructure underneath them.

bioinformatics↗

bioRxiv2024

FeatureForest: the power of foundation models, theusability of random forests

Once the work at the microscope is done, biological discoveries rely heavily on proper downstream analysis. This often amounts to first segmenting the biological objects of interest in the image before performing a quantitative analysis. Deep-learning (DL) is nowadays ubiquitous in such segmentation tasks. However, DL can be cumbersome to apply, as it often requires large amount of manual labeling to produce ground-truth data, and expert knowledge to train the models from scratch. Nonetheless, the performance of large foundation models, although trained on natural images, are improving on scientific images with every new model released. They, however, require either manual prompting or tedious post-processing to selectively segment the biological objects of interest. Classical machine learning algorithms, such as random forest classifiers, on the other hand, are well-established, easy to train, and often yield results of sufficient quality for downstream processing tasks, hence their continued popularity. Unfortunately, they are limited to objects with distinct, well-defined textures compared to their environment. This generally limits their usefulness to structures easy to recognize. Here, we present FeatureForest, an open-source tool that leverages the feature embeddings of large foundation models to train a random forest classifier, thereby providing users with a rapid way of semantically segmenting complex images using only a few labeling strokes. We demonstrate the improvement in performance over a variety of datasets, including large and complex volumetric electron microscopy stacks. Our implementation is available in napari, currently integrates four foundation models, and can easily be extended to any new model once they become available.

bioinformatics↗

bioRxiv2024

Hijacking and Integration of Algal Plastids andMitochondria in a Polar Planktonic Host

In oceanic plankton, various host organisms are capable of engulfing and temporarily integrating microalgae (photosymbiosis) or just their photosynthetic plastids (kleptoplastidy) as a solar-powered energy source. These cellular interactions can be considered to be representative of evolutionary steps in plastid acquisition in eukaryotes, but the underlying cellular mechanisms and dynamics are not fully understood. Here, we studied a kleptoplastidic dinoflagellate (RSD: Ross Sea Dinoflagellate), which is known to steal plastids of the microalga Phaeocystis antarctica. We tracked the morphology and activity of stolen plastids over several months by combining multimodal subcellular imaging and photophysiology. Upon integration inside a host vacuole, the volume of plastids and pyrenoids significantly increased and photosynthetic activity was boosted along with carbon fixation and transfer to the host. This may be supported by the retention of a 50-fold larger algal nucleus for [~]1 week. Once the algal nucleus was lost, there was a decrease in plastid volume and photosynthesis, but plastids were still beneficial for the host after > 2 months. Unlike other kleptoplastidic interactions, we showed that the algal mitochondrion was also stolen and retained for several months, transforming into an extensive network in close proximity with plastids. This highlights a new strategy in plankton along the continuum of plastid symbioses where both the energy-producing plastid and mitochondrion of a microalga are hijacked for several months by a host. This symbiosis that we found to be widely-distributed in polar regions suggests that plastid-mitochondrion interaction may have played a role in the evolution of plastid acquisition.

ecology↗

bioRxiv2020

Syndapin-2 mediates amyloid-β transcytosis at the brain endothelium: implications in Alzheimer's disease

A deficient transport of amyloid-(A{beta}) across the blood-brain barrier (BBB), and its diminished clearance from the brain, contributes to neurodegenerative and vascular pathologies, including Alzheimers (AD) and cerebral angiopathy, respectively. At the BBB, A{beta} efflux transport is associated with the low-density receptor-related protein 1 (LRP1). However, the precise mechanisms governing A{beta} transport across the BBB, in health and disease, remain to be fully understood. Recent evidence indicates that the LRP1 transcytosis occurs through a tubulation-mediated mechanism stabilised by syndapin-2. Here, we show that syndapin-2 is associated with A{beta} clearance via LRP1 across the BBB. We further demonstrate that risk factors for AD, A{beta} expression and ageing, are associated with a decline in the native expression of syndapin-2 within brain endothelium. Our data reveal that the syndapin-2-mediated pathway, and its balance with the endosomal sorting, are important for A{beta} clearance proposing a measure to evaluate AD and ageing, as well as a target for counteracting A{beta} build-up. Moreover, we provide evidence for the impact of the avidity of A{beta} assemblies in their trafficking across the brain endothelium and in LRP1 expression levels, which may affect the overall clearance of A{beta} across the BBB.

neuroscience↗

bioRxiv2020

Identification of amyloid beta oligomers in locus coeruleus (LC) neurons of Alzheimer's patients and their impact on LC oxidative stress, inhibitory neurotransmitter receptors and neuronal excitability

Amyloid {beta} oligomers (A{beta}O) are potent modulators of two key Alzheimers pathological processes, namely synaptic dysfunction and tau tangle formation in various brain regions. Remarkably, the impact of A{beta}O in one of the earliest brain regions to exhibit Alzheimers pathology, the locus coeruleus (LC), remains to be determined. Of particular importance is the effect of A{beta}O on the excitability of individual LC neurons. This parameter determines brain-wide noradrenaline (NA) release, and thus NA-mediated brain functions, including cognition, emotion and immune function, which are all severely compromised in Alzheimers. Using a mouse model of increased A{beta} production (APP-PSEN1), together with correlative histopathological analyses in post mortem Alzheimers patient samples, we determined the impact of A{beta} pathology on various correlates of LC neuronal integrity. A{beta}O immunoreactivity in the LC of APP-PSEN1 mice was replicated in patient samples, presenting as individual clusters located both intraneuronally, in mitochondrial compartments, as well as extracellularly in association with inhibitory synapses. No specific signal was detected in either patient control or wild type mouse samples. Accompanying this A{beta}O expression profile was LC neuronal hyperexcitability and indicators of oxidative stress in APP-PSEN1 mice. LC hyperexcitability arose from a diminished inhibitory effect of GABA, due to impaired expression and function of the GABA-A receptor (GABAAR) 3 subunit. Importantly, this altered LC 3-GABAAR expression profile overlapped with A{beta}O expression in both APP-PSEN1 mice and Alzheimers patient samples. Finally, strychnine-sensitive glycine receptors (GlyRs) remained resilient to A{beta}O-induced changes and their activation reversed LC hyperexcitability. Alongside this first demonstration of A{beta}O expression in the LC of Alzheimers patients, the study is also first to reveal a direct association between A{beta}O and LC neuronal excitability. GlyR-3-GABAAR modulation of A{beta}O-dependent LC hyperexcitability could delay the onset of cognitive and psychiatric symptoms arising from LC-NA deficits, thereby significantly diminishing the disease burden for Alzheimers patients.

neuroscience↗