Integrating representation learning, permutation, and optimization to detect lineage-related gene expression patterns

Nature Communications, Год журнала: 2025, Номер 16(1)

Опубликована: Янв. 27, 2025

Recent barcoding technologies allow reconstructing lineage trees while capturing paired single-cell RNA-sequencing (scRNA-seq) data. Such datasets provide opportunities to compare gene expression memory maintenance through branching and pinpoint critical genes in these processes. Here we develop Permutation, Optimization, Representation learning based single Cell Expression Lineage ANalysis (PORCELAN) identify lineage-informative or subtrees where are tightly coupled. We validate our method using synthetic data apply it recent scRNA-seq of lung cancer a mouse model embryogenesis C. elegans. Our pinpoints giving rise metastases new cell states, identified as most informative about overlap with known pathways involved progression. Furthermore, highlights differences how is maintained divisions embryogenesis, thereby providing tool for studying state across biological systems. the reconstruction cell's tree transcriptomic Here, authors present statistical automatically detect patterns related

Язык: Английский

---

Independent mechanisms of benzimidazole resistance across Caenorhabditis nematodes

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2025, Номер unknown

Опубликована: Март 15, 2025

Benzimidazoles (BZs), a widely used class of anthelmintic drugs, target beta-tubulin proteins, disrupt microtubule formation, and cause nematode death. In parasitic species, mutations in genes ( e.g. , isotype-1 beta-tubulin) are predicted to inhibit BZ binding associated with resistance. Similarly, the free-living Caenorhabditis elegans an ortholog, ben-1 primary drivers The recurrent association resistance beta-tubulins suggests that is repeatedly caused by genes, example repeated evolution drug across species. To evaluate hypothesis mediated beta-tubulin, we identified alleles wild strains from three species: C. briggsae tropicalis . We hypothesized that, if these species experienced similar selective pressures, they would evolve BZs any tbb-1 tbb-2 ). Using high-throughput development assays, tested found heterogeneous set variants were only two encode premature stop codon (W21stop) missense substitution (Q134H), but neither was might have evolved other or not Our findings reveal lack highlight importance defining mechanisms outside beta-tubulins.

Язык: Английский

---

Revealing a coherent cell state landscape across single cell datasets with CONCORD

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2025, Номер unknown

Опубликована: Март 15, 2025

Batch integration, denoising, and dimensionality reduction remain fundamental challenges in single-cell data analysis. While many machine learning tools aim to overcome these by engineering model architectures, we use a different strategy, building on the insight that optimized mini-batch sampling during training can profoundly influence outcomes. We present CONCORD, self-supervised approach implements unified, probabilistic scheme combining neighborhood-aware dataset-aware sampling: former enhancing resolution while latter removing batch effects. Using only minimalist one-hidden-layer neural network contrastive learning, CONCORD achieves state-of-the-art performance without relying deep auxiliary losses, or supervision. It generates high-resolution cell atlases seamlessly integrate across batches, technologies, species, prior assumptions about structure. The resulting latent representations are denoised, interpretable, biologically meaningful-capturing gene co-expression programs, resolving subtle cellular states, preserving both local geometric relationships global topological organization. demonstrate CONCORD's broad applicability diverse datasets, establishing it as general-purpose framework for high-fidelity of identity dynamics.

Язык: Английский

---

Neurogenesis in Caenorhabditis elegans

Genetics, Год журнала: 2024, Номер 228(2)

Опубликована: Авг. 21, 2024

Abstract Animals rely on their nervous systems to process sensory inputs, integrate these with internal signals, and produce behavioral outputs. This is enabled by the highly specialized morphologies functions of neurons. Neuronal cells share multiple structural physiological features, but they also come in a large diversity types or classes that give system its broad range plasticity. diversity, first recognized over century ago, spurred classification efforts based morphology, function, molecular criteria. Caenorhabditis elegans, precisely mapped at anatomical level, an extensive description most neurons, genetic amenability, has been prime model for understanding how neurons develop diversify mechanistic level. Here, we review gene regulatory mechanisms driving neurogenesis diversification neuron subclasses C. elegans. We discuss our current specification neuronal progenitors differentiation terms transcription factors involved ensuing changes expression chromatin landscape. The central theme emerged identity defined modules batteries are under control parallel yet interconnected mechanisms. focus how, achieve terminal identities, information along developmental lineages. Moreover, diversified postembryonically time-, sex-, activity-dependent manner. Finally, development can provide insights into evolution diversity.

Язык: Английский

---

Alternative splicing across theC. elegansnervous system

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Май 16, 2024

Abstract Alternative splicing is a key mechanism that shapes neuronal transcriptomes, helping to define identity and modulate function. Here, we present an atlas of alternative across the nervous system Caenorhabditis elegans . Our analysis identifies novel in genes such as unc-40 /DCC sax-3 /ROBO. Globally, delineate patterns differential almost 2,000 genes, estimate quarter undergo splicing. We introduce web interface for examination neuron types. explore relationship between type patterns, gene expression. identify RNA features correlate with splicing, describe enrichment microexons. Finally, compute regulatory network can be used generate hypotheses on regulation targets neurons.

Язык: Английский

---