# q-bio.GN Published articles for q-bio.GN. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Thalamus-Reasoning: A Multimodal Large Language Model for Multi-Sample Omics Reasoning DevFeed: [Thalamus-Reasoning: A Multimodal Large Language Model for Multi-Sample Omics Reasoning]() Original publisher: [Read original article]() Author: Maciej Sypetkowski, {\\L}ukasz Smoli\\'nski, Remigiusz Kinas, Joanna Krawczyk, Bart{\\l}omiej Hofman, Aleksandra Mo\\.zwi{\\l}{\\l}o, Micha{\\l} Pikusa, Tomasz Jetka, Rafa{\\l} Powalski Published: 2026-11-06T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Large Language Model](), [LLM evaluation / benchmarking](), [llm-reasoning](), [Fine-tuning]() Tags: [benchmark](), [best](), [compare](), [continuous](), [conversations](), [cs-ai](), [embedding](), [expression](), [geo](), [language-model](), [large-language-model](), [measurements](), [multimodal](), [performance](), [prediction](), [pretraining](), [q-bio-cb](), [q-bio-gn](), [reasoning](), [sample](), [supervised-fine-tuning](), [tokens](), [training]() ### AI overview Thalamus-Reasoning is a 9B multimodal model that represents bulk or single-cell transcriptomes as compact continuous embeddings, which can be combined with natural-language instructions and multiple samples in one context. Its training uses 20 billion tokens of continued pretraining and 5.21 million supervised conversations across 157 task groups, including examples linking measurements to biological annotations and outcomes. On the publication-disjoint GEO-OmicsQA benchmark, it leads the evaluated models on all six variants; it also leads on three of four held-out annotation tasks and has the best average PerturbQA performance among compared methods. Ablations favor continuous omics representations over text-only gene lists. These results are reported for the evaluations described in the abstract. ### Source excerpt arXiv:2605.06728v3 Announce Type: replace-cross Abstract: Large language models can reason over biological knowledge, but they are not naturally equipped to work with quantitative omics measurements. Existing approaches either convert expression profiles into long textual representations, often losing quantitative information, or connect omics encoders to language models for narrow tasks. We introduce Thalamus-Reasoning, a 9B multimodal language model that represents each bulk or single-cell transcriptome as a compact continuous embedding that can be interleaved with natural-language instructions and with other samples in the same context. A key contribution is a training corpus that connects experimental measurements with biological knowledge and task supervision. We combine templates linking transcriptomic profiles to annotations and experimental outcomes, reasoning examples connecting evidence across sources, and LLM-generated supervision with task-specific quality checks. The instruction-tuning corpus spans 157 biological and general-purpose task groups. Training includes continued pretraining on 20 billion tokens and supervised fine-tuning on 5.21 million conversations. We further introduce GEO-OmicsQA, a publication-disjoint benchmark of 3,000 questions requiring models to interpret and compare expression profiles from real GEO studies. Thalamus-Reasoning achieves the best results among the evaluated models on three of four held-out annotation tasks and on all six GEO-OmicsQA variants, and the best average performance among compared methods on PerturbQA perturbation-response prediction. Ablations show a clear advantage for continuous omics representations over a text-only gene-list representation. These results show that quantitative transcriptomic measurements and biological language can be integrated within a general conversational model, enabling both predictive tasks and language-guided interpretation of experimental samples. ## Hyper Input Convex Neural Networks for Shape Constrained Learning and Optimal Transport DevFeed: [Hyper Input Convex Neural Networks for Shape Constrained Learning and Optimal Transport]() Original publisher: [Read original article]() Author: Shayan Hundrieser, Insung Kong, Johannes Schmidt-Hieber Published: 2026-11-06T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Machine learning](), [Neural Network](), [Deep learning](), [mlp]() Tags: [abstract](), [apply](), [architecture](), [arxiv](), [based](), [convex](), [convex-functions](), [cs-lg](), [data](), [examples](), [existing](), [experiments](), [fewer](), [functions](), [input](), [learning](), [math-st](), [networks](), [neural](), [neural-networks](), [optimal-transport](), [q-bio-gn](), [stat-me](), [stat-ml](), [stat-th](), [transport]() ### AI overview HyCNNs combine Maxout networks with input convex neural networks (ICNNs) to build models that are always convex in their input and can use depth. The authors prove that HyCNNs need exponentially fewer parameters than ICNNs to approximate quadratic functions to a given precision. In synthetic convex regression and interpolation experiments, they report better predictive performance than ICNNs and MLPs. For high-dimensional optimal transport maps, including synthetic examples and single-cell data, HyCNNs often outperform ICNN-based methods and other baselines across varied settings. These are reported results; the supplied excerpt does not specify the settings or limitations. ### Source excerpt arXiv:2604.26942v2 Announce Type: replace Abstract: We introduce Hyper Input Convex Neural Networks (HyCNNs), a novel neural network architecture designed for learning convex functions. HyCNNs combine the principles of Maxout networks with input convex neural networks (ICNNs) to create a neural network that is always convex in the input, theoretically capable of leveraging depth, and performs reliably when trained at scale compared to ICNNs. Concretely, we prove that HyCNNs require exponentially fewer parameters than ICNNs to approximate quadratic functions up to a given precision. Throughout a series of synthetic experiments, we demonstrate that HyCNNs outperform existing ICNNs and MLPs in terms of predictive performance for convex regression and interpolation tasks. We further apply HyCNNs to learn high-dimensional optimal transport maps for synthetic examples and for single-cell data, where they oftentimes outperform ICNN-based neural optimal transport methods and other baselines across a wide range of setting. ## Genome-Factory: A Library for Tuning, Deploying, and Interpreting Genomic Foundation Models DevFeed: [Genome-Factory: A Library for Tuning, Deploying, and Interpreting Genomic Foundation Models]() Original publisher: [Read original article]() Author: Weimin Wu, Xuefeng Song, Yibo Wen, Qinjie Lin, Zhihan Zhou, Jerry Yao-Chieh Hu, Zhong Wang, Han Liu Published: 2026-11-06T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [AI Foundation Models](), [Fine-tuning](), [peft](), [Open Source Models & Datasets]() Tags: [automated](), [benchmarking](), [benchmarks](), [cs-lg](), [data](), [dna](), [download](), [embedding](), [fine-tuning](), [foundation-models](), [full](), [generation](), [genome](), [genomics](), [inference](), [interface](), [interpretability](), [library](), [python](), [q-bio-gn](), [sequences](), [workflow]() ### AI overview Genome-Factory combines data collection, model tuning, inference, benchmarking, and interpretability in a Python library for genomic foundation models. It automates sequence download and preprocessing; supports full and parameter-efficient fine-tuning across diverse models; and provides embedding extraction and DNA sequence generation. It includes two benchmarks and an interface for adding others, plus an open-source biological interpreter based on a sparse auto-encoder. The abstract reports validation of model and tuning compatibility, downstream performance on the two included benchmarks, and interpretation of DNABERT-2 representations. It does not provide specific performance results or details of the benchmark tasks. ### Source excerpt arXiv:2509.12266v3 Announce Type: replace-cross Abstract: We introduce Genome-Factory, the first integrated Python library for tuning, deploying, and interpreting genomic foundation models. Our core contribution is to simplify and unify the workflow for genomic model development: data collection, model tuning, inference, benchmarking, and interpretability. For data collection, Genome-Factory offers an automated pipeline to download genomic sequences and preprocess them. For model tuning, Genome-Factory supports both full and parameter-efficient fine-tuning across diverse genomic models. For inference, Genome-Factory enables both embedding extraction and DNA sequence generation. For benchmarking, we include two existing benchmarks and provide a flexible interface to incorporate additional benchmarks. For interpretability, Genome-Factory introduces an open-source biological interpreter based on a sparse auto-encoder. We validate the utility of Genome-Factory across three dimensions: (i) Compatibility with diverse models and fine-tuning methods; (ii) Benchmarking downstream performance using two open-source benchmarks; (iii) Biological interpretation of learned representations with DNABERT-2. These results highlight its practical value for real-world genomic analysis. GitHub: https://github.com/WeiminWu2000/Genome_Factory. ## Reclassification Is All You Missed: Surfacing Variant Reinterpretation Need in Online Health Communities at Population Scale DevFeed: [Reclassification Is All You Missed: Surfacing Variant Reinterpretation Need in Online Health Communities at Population Scale]() Original publisher: [Read original article]() Author: Kamalakkannan Ravi, Mengshu Nie, Annapurna Poduri, Cong Liu Published: 2026-11-06T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Machine learning]() Tags: [2014](), [abstract](), [advice](), [agent](), [agentic](), [all](), [arxiv](), [but](), [candidate](), [classification](), [code](), [communities](), [cs-ai](), [cs-cl](), [cs-cy](), [data](), [discuss](), [discussion](), [dry](), [health](), [need](), [online](), [q-bio-gn](), [scale]() ### AI overview PRISM uses an agentic large-language-model pipeline to screen public health-community posts for genetic variant reinterpretation needs, check later ClinVar classifications, and draft disclaimer-bearing outreach messages in a no-contact dry run. Across 3.9 million posts from 32 communities, relevant discussion was uncommon (0.288%) but more concentrated in specialist genetics communities than general medical-advice communities (6.30% versus 0.10%), and increased fivefold since 2014. The pipeline surfaced 123 candidate classification differences, including five shifts from uncertain to pathogenic or likely pathogenic. In 30 held-out drafts, a clinician judged 27 factually grounded but only 15 fit to send, showing that accuracy alone did not ensure acceptable outreach. The supplied text is an abstract/excerpt and does not describe the pipeline's detailed methods. ### Source excerpt arXiv:2611.07054v1 Announce Type: cross Abstract: Genetic variants are routinely reclassified in ClinVar after a patient is tested, sometimes from uncertain to pathogenic, yet patient recontact is inconsistent, with no clear duty and responsibility diffused among labs, providers, and patients. Many affected people discuss their results in public online health communities that clinical recontact often does not reach. We built PRISM, an agentic large-language-model pipeline that screens these posts, checks whether the variant was later reclassified, and drafts a disclaimer-bearing outreach message, in a no-contact dry run. Across 3.9 million posts from 32 communities, agent-predicted reinterpretation-relevant discussion is rare (0.288%) but concentrated in specialist genetics communities (6.30% versus 0.10% in general medical-advice communities) and growing fivefold since 2014; PRISM surfaces 123 candidate classification differences, 5 of them from an uncertain submission to pathogenic or likely pathogenic. On a held-out set of 30 real-post drafts, a clinician judged 27 factually grounded and 15 fit to send: factual accuracy is necessary but not sufficient for clinically acceptable outreach. PRISM contributes a population-scale reinterpretation-need map with code and data to be released upon publication. ## VenusGenome: Genome-Aware Masked Autoencoding for Microbial Genome Representation Learning DevFeed: [VenusGenome: Genome-Aware Masked Autoencoding for Microbial Genome Representation Learning]() Original publisher: [Read original article]() Author: Zhang Liang, Zhang Yukun, Tan Yang, Jiang Fan, Zhong Bozitao, Cheng Xiaoran, Hong Liang, Li Mingchen, Yu Yuanxi Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Representation learning](), [Machine learning](), [Model Development](), [Embeddings]() Tags: [1-7](), [abstract](), [also](), [analysis](), [arxiv](), [assembly](), [attention](), [autoencoder](), [bert](), [can](), [combines](), [context](), [contribution](), [cs-ai](), [cs-lg](), [defects](), [embedding](), [embeddings](), [encoding](), [fitness](), [genome](), [learning](), [q-bio-gn]() ### AI overview VenusGenome learns microbial genome representations by reconstructing masked gene embeddings from sparse genomic context. Its gene-level masked autoencoder encodes only visible genes and uses attention that accounts for contig boundaries, relative strand orientations, and gene positions, avoiding assumptions that separate contigs are adjacent or that reference strand labels alone convey biological orientation. It is pretrained on frozen protein embeddings from over 1.7 million isolate and metagenomic assemblies. In the reported evaluations, its frozen representations had the highest mean performance among evaluated models on nine of eleven primary metrics across genome-level phenotype prediction and gene-level tasks. Ablations favor genome-aware attention and masked autoencoding over BERT-style masked modeling, with further gains from higher masking ratios. An exploratory perturbation analysis links representation sensitivity to measured fitness defects; the abstract does not establish causation. ### Source excerpt arXiv:2611.05405v1 Announce Type: cross Abstract: Genomic-context models have shown that contextualizing protein embeddings can improve predictions of microbial function. We observe that a gene's protein embedding is partially predictable from the embeddings of neighboring genes in the genome, motivating high-ratio masking that requires reconstruction from sparser genomic context. Genomic modeling must also distinguish biological relationships from assembly conventions, since the ordering of separate contigs does not imply adjacency and strand labels depend on the reference orientation. We therefore introduce VenusGenome, a gene-resolution masked autoencoder that combines visible-only encoding with genome-aware attention incorporating contig boundaries, relative strand orientations, and gene positions. We pretrain this model on a curated corpus of over 1.7 million microbial genome assemblies from isolate and metagenomic sources, using frozen protein embeddings as gene-level inputs. Across genome-level phenotype prediction and gene-level tasks, the resulting frozen representations achieve the highest mean performance among the evaluated models on nine of eleven primary metrics. An exploratory gene-input perturbation analysis further links representation sensitivity to experimentally measured fitness defects. Ablations support the contribution of genome-aware attention and show the benefits of masked autoencoding over BERT-style masked modeling, with further gains from higher masking ratios. ## A Training-Free Socket for DNA in Frozen Language Models DevFeed: [A Training-Free Socket for DNA in Frozen Language Models]() Original publisher: [Read original article]() Author: Arash Fatehi Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [LLM Techniques](), [Large Language Model](), [linear-regression]() Tags: [cs-ai](), [dna](), [free](), [language](), [language-models](), [models](), [q-bio-gn](), [socket](), [training]() ### AI overview The socket transfers gene-level information from frozen Evo 2 DNA features into a frozen language model using one closed-form ridge regression, with no gradients through the host. On GeneHop, real DNA outperformed shuffled DNA across four hosts by 16.7-26.1 points; all results survived Holm correction, and null controls stayed within two points of sham. A text-only linear map let one plug serve other hosts at a test cost of at most 0.9 points. Nearest-neighbour retrieval captured 72-84% of the effect, while the plug scored 3.0-5.2 points higher on each host and retained most of its effect without training homologs. These results come from a single held-out test run under a frozen analysis plan. ### Source excerpt arXiv:2611.05401v1 Announce Type: cross Abstract: Connecting a DNA foundation model to a large language model (LLM) usually requires training a connector, and often the LLM too. We show that a closed-form map is enough for gene-level questions. Our method, the socket, fits one ridge regression (the plug) from frozen Evo 2 features to the host LLM's own representation of the gene's symbol, name, and description at an early layer. The plug's output replaces the hidden state at placeholder positions, and no gradient passes through the host. We test it on GeneHop, a five-way multiple-choice benchmark split by homology. In a single run on the held-out test set, under a frozen analysis plan, real DNA beats shuffled DNA on all four hosts: by +22.2 points [19.8, 25.0] on Qwen3.5-4B, +26.1 [24.3, 27.9] on Qwen3.6-27B, +16.7 [14.2, 19.7] on Mistral-7B, and +18.4 [16.2, 21.0] on OLMo-2-7B. All four survive Holm correction, every null control stays within two points of the sham level, and the development results agree within 2.3 points. A layer rule that uses no task items was selected on three hosts; it predicted the working layer of the fourth before any DNA-plug scoring on it. A text-only linear map lets one plug feed any other host, at a cost of at most 0.9 points on test. The socket conveys gene identity: nearest-neighbour retrieval recovers 72-84% of the effect, but the plug beats it on every host by 3.0-5.2 points and keeps most of its effect without training homologs. ## BEND 2.0: Benchmarking the Capabilities Needed to Design DNA DevFeed: [BEND 2.0: Benchmarking the Capabilities Needed to Design DNA]() Original publisher: [Read original article]() Author: Andreas Bjerregaard, Rachael DeVries, Dionysia Danai Brilli, Panagiotis Antoniadis, Yan Li, Anders Krogh, Wouter Boomsma, Ole Winther, Frederikke Isa Marin Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [AI Foundation Models](), [LLM evaluation / benchmarking](), [post-training]() Tags: [benchmarking](), [cs-lg](), [dna](), [foundation-models](), [framework](), [genomics](), [q-bio-gn](), [q-bio-qm]() ### AI overview BEND 2.0 evaluates DNA models across eight tasks and five experimental datasets, testing mutation-effect prediction and whether predictions can guide sequence design. Results depend on the model and representation access: linear readouts on frozen representations can match or outperform native outputs, while functional post-training improves capability. Readouts transfer across assays and genes, but average-effect accuracy can mask weaker context-specific predictions, and predicting combined mutation effects does not establish accurate interaction prediction. Model-guided design can reduce the number of candidates needed to find a desired response, yet context-specific and multi-cell objectives remain challenging. The benchmark finds signs of general utility, with performance still below saturation even on simple tasks. ### Source excerpt arXiv:2611.03806v1 Announce Type: cross Abstract: Genomic foundation models are increasingly used to propose DNA sequences and edits. But successful design requires capabilities that current benchmarks do not measure. Here we introduce BEND 2.0 (BENchmark for DNA language models), which tests how well models predict the effects of DNA changes and whether these predictions can guide sequence design. Across eight tasks built on five experimental datasets, we find that model choice and how representations are accessed both affect performance. Linear readouts fitted to frozen representations can match or exceed models' native outputs, and functional post-training substantially improves the models' capabilities. Linear readouts transfer well across experimental assays and genes, suggesting that models capture functional patterns beyond the data used to fit individual predictors. At the same time, strong prediction of average effects can conceal weaker prediction of context-specific effects, and accurate prediction of combined mutation effects does not imply accurate prediction of their interactions. In design tasks, model predictions can reduce the number of candidates needed to find a desired response, but still leave substantial room for better design choices, particularly for context-specific and multi-cell objectives. Our benchmark reveals that genomic models now display signs of general utility but have yet to saturate on performance in even simple tasks. ## Heterogeneous Graph Attention Network With Multimodal Feature Selection for Cancer Multiomics Integration DevFeed: [Heterogeneous Graph Attention Network With Multimodal Feature Selection for Cancer Multiomics Integration]() Original publisher: [Read original article]() Author: Sina Tabakhi, Charlotte Vandermeulen, Ian Sudbery, Haiping Lu Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Feature Engineering](), [Graphs]() Tags: [attention](), [biomarkers](), [cancer](), [cs-lg](), [cs-ma](), [deep-learning](), [diagnosis](), [feature](), [feature-selection](), [graph](), [integration](), [interpretability](), [machine-learning](), [multimodal](), [network](), [q-bio-bm](), [q-bio-gn]() ### AI overview HeteroGATomics combines multimodal feature selection with heterogeneous graph attention for cancer diagnosis. A multi-agent system builds feature- and patient-similarity networks for each omics modality, then combines them into one heterogeneous graph to learn omics-specific representations and integrate predictions. The abstract reports better results than baseline methods on most metrics across three cancer multiomics datasets, and better results than graph-based integration methods across metrics on an independent breast cancer cohort. It also reports identifying biomarkers that contribute to diagnosis outcomes. The supplied text does not provide dataset details, metric values, or implementation specifics. ### Source excerpt arXiv:2408.02845v2 Announce Type: replace Abstract: The increase in high-dimensional multiomics data demands advanced integration models to capture the complexity of human diseases. Graph-based deep learning integration models, despite their promise, struggle with small patient cohorts and high-dimensional features, often applying independent feature selection without modeling relationships among omics. Furthermore, conventional graph-based omics models focus on homogeneous graphs, lacking multiple types of nodes and edges to capture diverse structures. We introduce a Heterogeneous Graph ATtention network for omics integration (HeteroGATomics) to improve cancer diagnosis. HeteroGATomics performs multimodal feature selection through a multi-agent system, creating dedicated networks of feature and patient similarity for each omics modality. These networks are then combined into one heterogeneous graph for learning holistic omics-specific representations and integrating predictions across modalities. Experiments on three cancer multiomics datasets show that HeteroGATomics outperforms baseline methods across most evaluation metrics in cancer diagnosis. External validation on an independent breast cancer cohort further demonstrates its generalizability, where HeteroGATomics outperforms graph-based multiomics integration methods across evaluation metrics. Moreover, HeteroGATomics enhances interpretability by identifying important biomarkers contributing to the diagnosis outcomes. ## An auditable agentic pipeline for variant-to-gene prioritisation, validated against orthogonal functional evidence DevFeed: [An auditable agentic pipeline for variant-to-gene prioritisation, validated against orthogonal functional evidence]() Original publisher: [Read original article]() Author: Shakuntala Baichoo (Neel), Sohail (Neel), Sarkar, Bo Wang Published: 2026-11-03T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Agentic AI](), [llm-reasoning]() Tags: [agentic](), [benchmark](), [code](), [cs-ai](), [functional](), [language-model](), [model](), [overlap](), [pipeline](), [q-bio-gn](), [reranking](), [retrieval](), [study]() ### AI overview V2G records retrieved evidence alongside ranked candidate genes, but its reported language-model reranking gain appears to come from recalling known genes rather than interpreting that evidence. Among 135 held-out loci where retrieval succeeded and Open Targets L2G made predictions, reranking improved Precision@1 over the supplied evidence ordering (0.793 vs. 0.711; difference +0.081, 95% CI +0.038 to +0.129); no difference from Open Targets L2G or training-membership baselines was detected. Given only the trait and shuffled gene symbols, the model scored 0.864 on the same loci, and it named the reference gene in 59.4% of cases where V2G had not retrieved it. When gene symbols were moved to other genes' evidence rows, it followed the top-scored row and scored 0.284. Across 296 held-out loci, it kept the supplied top candidate in 86% of decisions; among natural retrieval cases, its departures selected a reference gene in 31 of 37 instances, displacing none. The study also reports more trait-matched mouse mutant or Mendelian phenotype annotations for its calls than for other candidates at the same locus, and provides a frozen benchmark, cached predictions, and reproduction code. ### Source excerpt arXiv:2611.01461v1 Announce Type: cross Abstract: Assigning GWAS variants to causal genes requires evidence from multiple resources. We present V2G, an agentic pipeline that records evidence retrieval and returns ranked genes alongside the retrieved evidence. We evaluate whether language-model reranking improves gene prioritisation and whether its gains depend on the retrieved evidence or recall of known associations. On 135 held-out loci where candidate retrieval succeeded and Open Targets L2G predicts, reranking by a language model improves Precision@1 over the evidence ordering it receives, 0.793 against 0.711 (+0.081, 95% CI +0.038 to +0.129), and over nearest-gene assignment (+0.178); no difference from Open Targets L2G or from training-membership baselines was detected. Given only the trait and the candidate gene symbols, in shuffled order, the same model scores 0.864 on these loci (three seeds, 0.859 - 0.867) and names the reference gene at 0.594 of the loci where the pipeline never retrieved it; with each gene's symbol moved to another gene's evidence row it follows the top-scored row and scores 0.284. The gain over the evidence ordering is recall of known genes, and the evidence as presented anchors the model to the ordering rather than informing it. Across the 296 held-out loci (five seeds each) the model keeps the given top candidate at 86% of decisions; at naturally retrieved loci its departures select a reference gene in 85% of departure decisions (31 of 37 loci), and none displaced one. Its calls carry trait-matched mouse mutant or Mendelian phenotype annotations more often than other candidates at the same locus. Train-test overlap is reported throughout. The study includes a frozen benchmark, cached predictions and reproduction code. ## HistoTx: Early fusion of H&E images and spatial transcriptomics at varying spatial transcriptomics resolution with self-supervised learning DevFeed: [HistoTx: Early fusion of H&E images and spatial transcriptomics at varying spatial transcriptomics resolution with self-supervised learning]() Original publisher: [Read original article]() Author: Alice Driessen, Emil Malmsten, Steffen Wolf, Bernd Illing, Isinsu Katircioglu, Adriano Martinelli, Rolf Stahel, Solange Peters, Enriqueta Felip, Sanjay Popat, Krisztian Homicsko, Raphael Gottardo, Mar Published: 2026-11-03T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence](), [arXiv: Machine Learning]() Topics: [Machine learning](), [multimodal-ai](), [Self-supervised learning](), [Transformers]() Tags: [architecture](), [cs-ai](), [cs-lg](), [data](), [deep](), [expression](), [foundation-model](), [fusion](), [images](), [inference](), [input](), [interaction](), [learning](), [modal](), [model](), [multimodal](), [patch](), [performance](), [prediction](), [pretraining](), [q-bio-gn](), [q-bio-qm](), [self-supervised-learning](), [spatial]() ### AI overview HistoTx combines H&E image patches and spatial transcript tokens at arbitrary spatial resolutions in a shared vision transformer. It is trained by continuing self-supervised pretraining of a vision-only pathology foundation model on paired image and spatial transcriptomics data. The abstract reports that image-only inference performs on par with the baseline on most tasks and improves tasks that benefit from molecular knowledge, such as gene expression prediction. When transcript inputs are also available at inference, the joint model outperforms either modality alone. ### Source excerpt arXiv:2611.02681v1 Announce Type: new Abstract: Spatial transcriptomics maps gene expression across tissue sections and, when integrated with whole slide images, jointly captures both morphological architecture and molecular activity to provide a rich, multimodal view of tissue biology. ML approaches combining spatial transcriptomics and WSIs typically learn a mapping from morphological features to molecular profiles via contrastive alignment or supervised fine-tuning of pathology foundation models. In both cases, transcriptomics serves as a training signal rather than a true input modality, leaving inference fundamentally image-driven. We propose an early-fusion vision transformer architecture in which transcript tokens are constructed at arbitrary spatial resolution, merged with image patch tokens, and jointly processed through a shared transformer stack for deep cross-modal interaction. Starting from a pretrained vision-only pathology foundation model, we train HistoTx by continuing self-supervised pretraining on paired image and spatial transcriptomics data and demonstrate that: 1) at inference with image only, performance remains on par with the baseline on most tasks while improving on tasks that inherently benefit from molecular knowledge, such as gene expression prediction; 2) when both modalities are available at inference, jointly providing image and transcript inputs outperforms either modality alone. ## Locality-Controlled Out-of-Distribution Guidance for Robust Regulatory DNA Sequence Design DevFeed: [Locality-Controlled Out-of-Distribution Guidance for Robust Regulatory DNA Sequence Design]() Original publisher: [Read original article]() Author: Dabin Jeong, Marwane Bourdim, Mo Lotfollahi Published: 2026-11-02T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Enhancing AI model performance](), [Fine-tuning]() Tags: [abstract](), [all](), [applying](), [arxiv](), [background](), [based](), [behavior](), [but](), [can](), [challenge](), [code](), [control](), [cs-lg](), [design](), [distribution](), [dna](), [guidance](), [locality](), [q-bio-gn](), [regulatory](), [robust](), [sequence]() ### AI overview LOCOGEN aims to generate regulatory DNA sequences that balance predicted activity, novelty, and plausibility, addressing a gap the authors say single-reward evaluation can hide through reward hacking. It treats guidance as position-dependent: sequence-level invariance learning identifies a motif mask whose positions are protected, while out-of-distribution damping is applied only to background positions, where more exploration is allowed. The method runs during sampling on a frozen diffusion backbone without retraining it. The abstract reports that LOCOGEN reaches the previously unmet property-novelty-plausibility region across HepG2, K562, and SK-N-SH; it provides no quantitative results in this excerpt. ### Source excerpt arXiv:2610.41736v1 Announce Type: new Abstract: Designing regulatory DNA sequences for targeted gene-expression control is an important challenge. Recent approaches frame this task as property-conditioned generation, where diffusion models trained on natural sequences are guided toward high predicted activity through reward-based fine-tuning or sampling-time guidance. We show that single-reward evaluation can obscure reward-hacking behavior. By decomposing evaluation into property, novelty, and plausibility, we identify an empirically underexplored region of the evaluation space in which all three objectives are satisfied, but no existing baseline succeeds. We postulate that this gap arises because prior methods treat guidance primarily as a matter of magnitude. In regulatory DNA, however, sequence grammar is spatially nonuniform: motif positions are sensitive to perturbation, while background regions allow more exploration. We therefore propose LOCOGEN by reformulating guidance as a problem of locality. LOCOGEN learns an invariant motif mask through sequence-level invariance learning to protect motif positions, while applying OOD damping only to background positions. It operates at sampling time on a frozen diffusion backbone, without retraining the backbone. Across HepG2, K562, and SK-N-SH, LOCOGEN reaches this previously underexplored region of the property-novelty-plausibility space. Code is available at https://github.com/Lotfollahi-lab/LocoGen. ## Localizing predictive concepts encoded in neural representations via hierarchical testing DevFeed: [Localizing predictive concepts encoded in neural representations via hierarchical testing]() Original publisher: [Read original article]() Author: Viktoriia Huryn, Marco Simnacher, Jim Berend, Wojciech Samek, Christoph Lippert, Sonja Greven, Uwe Ohler Published: 2026-11-02T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Neural Network](), [clustering](), [Representation learning](), [Predictive AI]() Tags: [cluster](), [clusters](), [concepts](), [conditional](), [control](), [cs-lg](), [deep-learning](), [dna](), [error-control](), [information](), [model](), [network](), [neural](), [prediction](), [q-bio-gn](), [regulatory](), [sequence](), [simulation](), [testing]() ### AI overview CUTTIE identifies groups of neural units that carry predictive information while controlling the family-wise error rate. It clusters each layer's units into a tree, tests clusters top-down with conditional independence tests, and uses conditional attribution maps to reveal the motifs encoded by units in selected clusters. In a controlled simulation, it identified clusters whose concepts affect the target. Applied to a convolutional model predicting cell-type-specific chromatin accessibility from DNA sequence, its selected clusters mapped to known regulatory motifs. The abstract positions this approach against exploratory ranking and visualization without population-level statistical guarantees, and testing methods that do not control error across many unit sets. ### Source excerpt arXiv:2610.41046v1 Announce Type: new Abstract: Understanding which regulatory sequence motifs a genomic deep learning model has learned, with statistical error guarantees, is essential for trusting its use in downstream tasks such as variant effect prediction. Existing approaches to identifying which units in a trained network encode human- understandable concepts fall into two groups. Exploratory methods rank and visualize units but offer no statistical guarantee that the selected units are associated with the prediction target across the population of inputs. Testing-based methods test only single units or one pre-defined set of them, and do not control the error across the many sets of units. We introduce CUTTIE (Cluster-tree Unit Top-down Testing of Independence with Error control), which clusters the units of each layer into a tree and tests the clusters top-down with conditional independence tests, localizing information about the target while controlling the family-wise error rate. Thereby, we identify clusters that carry information about the target beyond the rest of the layer. For the units in the identified clusters, conditional attribution maps then reveal what motifs they encode. In a controlled simulation, the procedure identifies the clusters whose concepts affect the target. We apply the procedure to a convolutional model predicting cell-type-specific chromatin accessibility from DNA sequence, where the concepts encoded in the identified clusters map onto known regulatory motifs. ## GenoBERT: A Language Model for Accurate Genotype Imputation DevFeed: [GenoBERT: A Language Model for Accurate Genotype Imputation]() Original publisher: [Read original article]() Author: Lei Huang, Chuan Qiu, Kuan-Jui Su, Anqi Liu, Yun Gong, Weiqiang Lin, Lindong Jiang, Chen Zhao, Meng Song, Jeffrey Deng, Qing Tian, Zhe Luo, Ping Gong, Hui Shen, Chaoyang Zhang, Hong-Wen Deng Published: 2026-10-30T04:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [bert]() Tags: [attention](), [benchmarking](), [bias](), [correlation](), [cs-ai](), [cs-lg](), [datasets](), [foundation](), [framework](), [genomics](), [language-model](), [panel](), [performance](), [q-bio-gn](), [range](), [resilience](), [sparsity](), [sufficient](), [transformer]() ### AI overview GenoBERT imputes missing phased genotypes without a reference panel, using a transformer with self-attention to capture short- and long-range linkage disequilibrium. In tests on the Louisiana Osteoporosis Study and 1000 Genomes Project across ancestry groups and 5-50% missingness, it had the highest overall accuracy among the four named baselines: Beagle5.4, SCDA, BiU-Net, and STICI. The abstract reports r² around 0.98 with up to 25% missing and above 0.90 at 50% missing. It also reports consistent gains across ancestries and resilience to small samples and weak LD. LD-decay analyses supported a 128-SNP context window (about 100 Kb) for local correlations. These findings are based on the abstract; it does not provide detailed methods or per-group results. ### Source excerpt arXiv:2604.00058v2 Announce Type: replace-cross Abstract: Genotype imputation enables dense variant coverage for genome-wide association and risk-prediction studies, yet conventional reference-panel methods remain limited by ancestry bias and reduced rare-variant accuracy. We present Genotype Bidirectional Encoder Representations from Transformers (GenoBERT), a transformer-based, reference-free framework that tokenizes phased genotypes and uses a self-attention mechanism to capture both short- and long-range linkage disequilibrium (LD) dependencies. Benchmarking on two independent datasets including the Louisiana Osteoporosis Study (LOS) and the 1000 Genomes Project (1KGP) across ancestry groups and multiple genotype missingness levels (5-50%) shows that GenoBERT achieves the highest overall accuracy compared to four baseline methods (Beagle5.4, SCDA, BiU-Net, and STICI). At practical sparsity levels (up to 25% missing), GenoBERT attains high overall imputation accuracy ($r^2 approx 0.98$) across datasets, and maintains robust performance ($r^2 > 0.90$) even at 50% missingness. Experimental results across different ancestries confirm consistent gains across datasets, with resilience to small sample sizes and weak LD. A 128-SNP (single-nucleotide polymorphism) context window (approximately 100 Kb) is validated through LD-decay analyses as sufficient to capture local correlation structures. By eliminating reference-panel dependence while preserving high accuracy, GenoBERT provides a scalable and robust solution for genotype imputation and a foundation for downstream genomic modeling. ## Advancing Clinical Genomic Analysis with a Skill-Centred Agent DevFeed: [Advancing Clinical Genomic Analysis with a Skill-Centred Agent]() Original publisher: [Read original article]() Author: Qiaoyu Zheng, Weike Zhao, Yanjie Fan, Ye Cao, Mengqi Jiang, Xiao Zhou, Xiaoman Zhang, Pengcheng Qiu, Yongguo Yu, Yanfeng Wang, Ya Zhang, Chaoyi Wu, Weidi Xie Published: 2026-10-30T04:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Artificial Intelligence](), [decision-making](), [Agent Skills](), [human in the loop agent]() Tags: [agent](), [ai](), [ai-collaboration](), [analysis](), [benchmark](), [cases](), [collaboration](), [cs-ai](), [databases](), [decision-making](), [deepseek](), [dependency](), [disease](), [domain](), [experts](), [feedback](), [guidelines](), [human-ai-collaboration](), [integrate](), [labor](), [language](), [q-bio-gn]() ### AI overview ClinGeX combines 105 curated executable skills using task-specific retrieval and dependency-aware orchestration to support clinical genomic analysis. On 223 temporally held-out cases, it reached 71.7% macro-averaged task-level concordance with expert classifications, versus 45.7% for the strongest comparator. In a counterbalanced study with 45 professionals assessing 135 cases, ClinGeX assistance raised raw case-level concordance in variant analysis, rare-disease interpretation, and molecular oncology, and reduced reported interpretation time across scenarios. An exploratory, non-interventional one-month assessment involved 11 professionals reporting on 96 self-selected episodes; they reported frequent adoption of conclusions and time savings. These findings suggest potential to augment specialist interpretation, while the real-world evidence is limited by its exploratory design and self-selected cases. ### Source excerpt arXiv:2610.40556v1 Announce Type: new Abstract: Clinical genomic analysis requires experts to integrate evolving guidelines, databases, and computational evidence. As sequencing has become increasingly accessible, this complex and labor-intensive process has become a major bottleneck to the timely use of genomic information in clinical decision-making. To fill this gap, we developed ClinGeX, a large-language-model agent that composes 105 curated executable skills through task-specific retrieval and dependency-aware orchestration. On 223 temporally held-out benchmark cases, ClinGeX achieved 71.7\% macro-averaged task-level concordance with expert reference classifications, compared with 45.7\% for the strongest comparator. In a three-condition, counterbalanced human-AI collaboration study, 45 clinical genomic professionals assessed 135 clinically derived cases under routine (no AI assistance), DeepSeek-assisted and ClinGeX-assisted conditions. ClinGeX assistance increased raw case-level concordance across criterion-level variant analysis, genomic interpretation in rare disease and molecular oncology decision-making (51.1\%, 57.8\%, 64.4\% versus 15.6\%, 31.1\%, 20.0\% in comparator conditions) and reduced reported interpretation time in all scenarios. During an exploratory one-month real-world non-interventional assessment, 11 experienced professionals submitted feedback on 96 self-selected analysis episodes and reported frequent conclusion adoption and time savings. These findings support executable domain skills as a promising approach to augment specialist genomic interpretation. ## Adaptive Allele-Haplotype Incidence Indexing for Pangenome Queries: Representation, Reconstruction, and Benchmarking DevFeed: [Adaptive Allele-Haplotype Incidence Indexing for Pangenome Queries: Representation, Reconstruction, and Benchmarking]() Original publisher: [Read original article]() Author: Roberto Garrone Published: 2026-10-30T04:00:00Z Content type: article Language: en Sources: [arXiv: Databases]() Topics: [Compression](), [Encoding](), [Structured-data](), [Phoronix Test Suite](), [Graph Mining]() Tags: [22](), [abstract](), [adaptive](), [analysis](), [archive](), [arxiv](), [assertions](), [benchmarking](), [but](), [canonical](), [carrier](), [catalog](), [comparison](), [compression](), [conditional](), [context](), [cs-db](), [cs-ds](), [custom](), [edge](), [indexing](), [q-bio-gn](), [queries]() ### AI overview The study tests two representations for phased-variant queries: H1 stores adaptive allele-haplotype incidence rows, while H2 encodes ordered edge paths from a fixed, nonoverlapping variant catalog. In a chromosome 1 interval, H1's adaptive files used 0.872 MiB (0.207 MiB after whole-file Zstandard compression), compared with 1.126 MiB for a custom Roaring-like layout and 1.190 MiB for a bitmap-payload baseline. For a separate 22,727-site SNV projection, context-Huffman coding reduced H2 edge sequences from 75,130,882 to 5,156,418 bytes. Notebook carrier-equivalence and graph-adjacency checks passed, and reconstruction is established conditionally for abstract H2 walks. The results show restricted implementation feasibility; cross-callset phasing, nucleotide-level structural-variant reconstruction, and comparison with production indexes remain unresolved, so they do not establish superior end-to-end pangenome compression. ### Source excerpt arXiv:2512.21320v3 Announce Type: replace-cross Abstract: Population-scale analysis of phased genetic variants requires both allele-carrier retrieval and ordered haplotype context. We examine H1, an adaptive allele-haplotype incidence-row index, and H2, a canonical edge-path representation derived from a fixed, nonoverlapping variant catalog. We prove conditional reconstruction of abstract H2 walks and distinguish this property from unrestricted equivalence to genomic variation graphs. A completed local notebook experiment used a chromosome 1 interval (20-22 Mb), 200 diploid individuals, 24,921 SNV/indel records, and 45 SVTYPE-annotated records. For 24,966 extracted nonreference rows across 400 haplotypes, the serialized H1 incidence-row files occupied 0.872 MiB with a simple adaptive layout and 1.126 MiB with a custom Roaring-like layout, versus 1.190 MiB for a bitmap-payload baseline; a whole-file Zstandard archive of the adaptive layout occupied 0.207 MiB. For a separately constructed nonoverlapping SNV projection containing 22,727 sites, indexed H2 edge sequences occupied 75,130,882 bytes in a raw format and 5,156,418 bytes with context-Huffman encoding (14.57-fold reduction). Carrier-equivalence and graph-adjacency assertions passed in the notebook, but full crosscallset phasing, nucleotide-level SV reconstruction, and comparison with production indexes remain unresolved. The measurements establish a restricted implementation feasibility result, not superior end-to-end pangenome compression. ## Evaluating and improving AI agents for sequencing protocol understanding DevFeed: [Evaluating and improving AI agents for sequencing protocol understanding]() Original publisher: [Read original article]() Author: Chi-Lam Poon, Michael Goneos, Zezheng Song, Xi Chen Published: 2026-10-30T04:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Artificial Intelligence](), [llm-reasoning](), [gaia]() Tags: [agents](), [ai](), [ai-agents](), [benchmark](), [benchmarks](), [complex](), [connectivity](), [cross](), [cs-ai](), [experimental](), [experiments](), [family](), [frontier](), [generation](), [improvement](), [knowledge](), [language](), [language-model](), [library](), [model](), [protocol](), [q-bio-gn]() ### AI overview LibStructBench evaluates whether AI agents can reconstruct sequencing-library structures and molecular workflows from protocol documents. Frontier agents often get close but remain unreliable on exact reconstruction, especially molecular-state representation and workflow connectivity. Changing the underlying language model caused larger average performance differences than changing the agent harness, but only molecular-state reconstruction showed a statistically significant difference. In a small pilot, added knowledge improved oligo-family recovery without improving mean molecular-workflow scores. Accumulating procedural rules and cross-protocol examples produced non-monotonic validation results and no mean held-out workflow improvement along the tested trajectory. ### Source excerpt arXiv:2610.37851v1 Announce Type: new Abstract: Existing scientific-agent benchmarks rarely evaluate whether agents can reconstruct the exact molecular structures and generation workflows specified by sequencing protocols, a capability critical for analyzing customized sequencing assays. We introduce LibStructBench, a document-grounded benchmark for reconstructing sequencing-library structures and molecular workflows. Our experiments show that frontier agents often produce approximately correct reconstructions but remain unreliable at exact reconstruction, particularly in molecular-state representation and workflow connectivity. Switching the underlying language model produced larger average performance differences than switching agent harnesses, with a statistically significant difference only for molecular-state reconstruction. Additional knowledge improved oligo-family recovery in a small pilot without improving mean molecular-workflow scores. Autonomous accumulation of procedural rules and cross-protocol examples showed non-monotonic validation performance and no mean held-out molecular-workflow improvement in the tested trajectory. These findings highlight the need for more controlled and verifiable approaches to scientific reasoning about molecular processes in complex experimental protocols. ## Enhancing Clinical Classifier Robustness with Deep Representation Learning Based OOD Detection on DNA Methylation Data DevFeed: [Enhancing Clinical Classifier Robustness with Deep Representation Learning Based OOD Detection on DNA Methylation Data]() Original publisher: [Read original article]() Author: Md Zahangir Alom, Quynh T. Tran, Breuer Alexandar, Brent A. Orr Published: 2026-10-28T04:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence](), [arXiv: Machine Learning]() Topics: [Representation learning](), [Deep learning](), [Deep neural networks](), [Embeddings](), [Machine learning]() Tags: [abstract](), [analysis](), [approach](), [array](), [arxiv](), [based](), [brain](), [bt](), [but](), [can](), [classification](), [cs-ai](), [cs-lg](), [data](), [deep](), [detection](), [dna](), [enhancing](), [learning](), [ood](), [q-bio-gn](), [robustness](), [with]() ### AI overview The proposed framework uses deep neural network bottleneck embeddings to flag tumor methylation samples that fall outside the training distribution, offering a potential quality-control layer for classification. It compares five detection methods across 200- and 500-dimensional embeddings; 1D-SSA performed best. In sarcoma-trained/brain-tumor-OOD evaluation, it reached AUROC 0.975 and TNR@95TPR 0.930; in the reverse setting, AUROC was 0.9569 and TNR@95TPR 0.8660. On an independent clinical test set, reported mean AUROC was 0.976, FPR@95TPR 0.065, TNR@95TPR 0.935, and DE 0.032. The supplied text is an abstract and does not provide implementation details. ### Source excerpt arXiv:2610.36344v1 Announce Type: cross Abstract: DNA methylation (DNAM) array-based tumor classification is a powerful clinical diagnostic tool, but models trained on fixed reference cohorts can produce faulty predictions for tumor types not represented in the training distribution. We propose an end-to-end deep representation learning-based out-of-distribution (OOD) detection framework for DNAm array data. Deep neural network (DNN) classifiers were trained on public sarcoma (SARC) (GSE140686) and brain tumor (BT) (GSE90496) cohorts, and feature vectors were extracted from bottleneck layers. Five OOD methods, including 1dimension (1D)-Subspace Analysis (1D-SSA), K-Nearest Neighbors (KNN), Ridge Regression, Mahalanobis Distance (MD), and Sum of Squared Distances (SSD), were explored with 200 and 500-dimensional embeddings. Spectral discrepancy was used to discriminate in-distribution (IND) from OOD samples. Performance was measured via cross-validation, holdout samples, and an independent clinical dataset. 1D-SSA was the most effective approach. In the sarcoma-IND/brain tumor-OOD setting, the framework achieved an AUROC of 0.975 and True Negative Rate (TNR)@95TPR of 0.930, and an AUROC of 0.9569 and TNR@95TPR of 0.8660 for the brain tumor-IND/sarcoma-OOD setting. On the independent clinical test set, the method yielded a mean AUROC of 0.976, a DE of 0.032, an FPR@95TPR of 0.065, and a TNR@95TPR of 0.935. These results show deep representation-based OOD detection as a practical quality-control layer for clinical-grade methylation-based tumor classification systems. ## scPyviewer: a Python-native interactive viewer from AnnData single-cell data DevFeed: [scPyviewer: a Python-native interactive viewer from AnnData single-cell data]() Original publisher: [Read original article]() Author: Hao Xuan, Yu Huang, Jiang Bian, Xiangtao Liu Published: 2026-10-28T04:00:00Z Content type: release Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Streamlit](), [pandas]() Tags: [api](), [benchmark](), [benchmarks](), [cells](), [comparison](), [cs-ai](), [dataset](), [datasets](), [embeddings](), [explore](), [expression](), [faster](), [framework](), [implementation](), [infrastructure](), [interaction](), [interactive](), [interactive-viewer](), [lightweight](), [native](), [python](), [q-bio-gn]() ### AI overview scPyviewer provides a Python-native interactive viewer that reads AnnData directly, avoiding conversion to Seurat for sharing single-cell results. In an audit against three R Shiny tools, it matched or exceeded baseline capabilities (7/7), with native AnnData ingestion and cross-dataset comparison over shared genes and matched cell types as distinct features. Across three datasets of 22,315 to about 313,000 cells, it rendered every core view faster (up to 3.6x) and used less memory (5.2x lower on the smallest dataset). At 313,000 cells, it completed all views while Seurat exceeded an 8 GB memory limit and failed. Its Python API also returns Matplotlib figures and pandas tables. These findings are reported in the supplied abstract. ### Source excerpt arXiv:2610.35890v1 Announce Type: cross Abstract: Motivation: Interactive web tools enabling non-programmers to explore single-cell datasets, embeddings, gene expression, metadata, and marker genes are standard laboratory infrastructure. However, actively maintained tools (e.g., ShinyCell, ScRDAVis, sCIRCLE) rely on R Shiny and require Seurat objects. Python/scanpy, the dominant framework for single-cell RNA-seq and spatial analysis, lacks a lightweight, native option pairing a shareable viewer with a scriptable Python API. Sharing scanpy results currently requires converting to Seurat or sharing code notebooks. Results: We present scPyviewer, a web viewer that ingests AnnData objects directly and reproduces incumbent interaction patterns within the Python stack. In a feature-parity audit against three R Shiny tools, scPyviewer matches or exceeds baseline capabilities (7/7), uniquely offering native AnnData ingestion without conversion and cross-dataset comparison over shared genes and matched cell types. Benchmark tests across three datasets (22,315 to ~313,000 cells) show scPyviewer renders every core view faster (up to 3.6x speedup) using lower memory (5.2x lower on the smallest dataset). On 313,000 cells, scPyviewer completes all views while Seurat exhausts an 8 GB memory limit and fails. Beyond the app, scPyviewer exposes a Python API returning Matplotlib figures and pandas tables for publication-ready output. Availability and implementation: Implemented in Python 3.11 (scanpy 1.11.5, anndata 0.12.19, streamlit 1.59.2, plotly 6.9.0) with a one-command interface to reproduce benchmarks, figures, and launch the app. Source code: https://github.com/xuan13hao/scPyviewer.git ## STRAND: Sequence-Conditioned Transport for Single-Cell Perturbations DevFeed: [STRAND: Sequence-Conditioned Transport for Single-Cell Perturbations]() Original publisher: [Read original article]() Author: Boyang Fu, Sameer Gabbita, George Dasoulas, Xiang Lin, Shanghua Gao, Xiaorui Su, Soumya Ghosh, Marinka Zitnik Published: 2026-10-28T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [few-shot](), [Language models]() Tags: [ablation](), [benchmarks](), [best](), [cells](), [conditional](), [control](), [cs-lg](), [discrimination](), [dna](), [element](), [elements](), [encoding](), [few-shot](), [generative](), [identifier](), [model](), [modeling](), [models](), [prediction](), [q-bio-cb](), [q-bio-gn](), [sequence](), [transport]() ### AI overview STRAND predicts cellular responses to genetic perturbations by encoding the DNA sequence at the targeted locus and conditioning a transport process from control to perturbed single-cell states. This allows locus-specific inputs and zero-shot predictions for targets absent from training, addressing the fact that different loci in one gene can produce distinct responses. In CRISPRi screens across K562, Jurkat, and RPE1 cells, it achieved the best average rank on few-shot and unseen-gene benchmarks, with up to 33% gains in perturbation discrimination and up to 0.12 improvement in ΔPearson when transferring to held-out cell lines. Ablations attributed the gains to sequence conditioning and conditional transport. The supplied text is an abstract and does not provide further evaluation details. ### Source excerpt arXiv:2602.10156v2 Announce Type: replace-cross Abstract: Prediction of the transcriptional response of a cell to a genetic perturbation is a central problem in the computational modeling of gene regulation. Perturbations of the same gene can produce different transcriptional responses depending on the targeted locus, such as an alternative transcription start site or a distal regulatory element, and gene-level perturbation models represent these distinct interventions with a single gene identifier. We introduce STRAND, a generative model that encodes the regulatory DNA sequence at the perturbation locus and uses this encoding to parameterize a conditional transport process from control to perturbed single-cell states. Sequence conditioning enables zero-shot prediction for perturbation targets absent from training and supports locus-resolved inputs at 128 bp model resolution within a 524 kb sequence window. On CRISPRi screens in K562, Jurkat, and RPE1 cells, STRAND achieves the best average rank on few-shot and unseen-gene benchmarks, with gains of up to 33% in perturbation discrimination, and improves transfer to held-out cell lines by up to 0.12 in $\Delta$Pearson. Ablation studies attribute these gains to sequence conditioning and to conditional transport. At the locus level, STRAND distinguishes alternative transcription start sites and validated regulatory elements that gene-identifier models represent as a single perturbation. STRAND is available at https://github.com/mims-harvard/STRAND. ## HELICS: Biobank-scale Conditional Synthetic Genome Generation via Latent Flow Matching DevFeed: [HELICS: Biobank-scale Conditional Synthetic Genome Generation via Latent Flow Matching]() Original publisher: [Read original article]() Author: Ahmad Abdel-Azim, Xihong Lin Published: 2026-10-28T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Neural Network](), [Dataset development]() Tags: [conditional](), [cs-lg](), [flow-matching](), [framework](), [generation](), [genome](), [q-bio-gn](), [synthetic]() ### AI overview HELICS generates conditional synthetic genome-wide cohorts by compressing more than 600,000 SNPs into 4,096 continuous tokens with chromosome-level convolutional autoencoders, then modeling the combined representation with a transformer-based conditional flow matching model. Trained on more than 417,000 UK Biobank genomes, it generates common-variant cohorts without haplotype phasing. The abstract reports stronger genotype fidelity than existing haplotype-based simulators, with high external allele-frequency agreement and better preservation of local and long-range SNP correlations. Conditioning on ancestry and genetic-liability summaries supports targeted generation; perturbation analyses found that increasing conditioned disease liability produced SNP dosage changes correlated with published association weights. These findings are reported in the supplied abstract; it does not provide further methodological or evaluation details. ### Source excerpt arXiv:2610.36361v1 Announce Type: cross Abstract: Individual-level genotype data underpin genetic discovery, disease-risk modeling, and therapeutic development, yet many critical settings remain data-limited, including underrepresented ancestries, rare or imbalanced disease cohorts, and restricted-access biobanks. Conditional synthetic genome generation would support privacy-aware benchmarking, simulation, and data augmentation, but requires learning an ultra-high-dimensional discrete distribution over hundreds of thousands to tens of millions of correlated single-nucleotide polymorphisms (SNPs), or sites of DNA variation. We introduce HELICS: High-fidelity gEneration via Latent Interpolant Conditional Synthesis, a phasing-free latent flow matching framework for genome-wide SNP-level generation from biobank-scale genotype data. HELICS tokenizes chromosomes with convolutional autoencoders, reducing >600K SNPs to 4,096 continuous latent tokens, and trains a transformer parameterized conditional flow matching model over the concatenated multi-chromosome latent representation. Trained on over 417K genomes in the UK Biobank, HELICS generates genome-wide common-variant synthetic cohorts without haplotype phasing and achieves stronger fidelity to real genotypes than existing haplotype-based simulators, including high allele-frequency agreement in external evaluation and superior preservation of local and long-range SNP correlation structure. Conditioning on ancestry and genetic liability summaries enables targeted generation across genetic risk profiles; in perturbation analyses, by increasing the conditioned disease trait liability, HELICS induces SNP-level dosage changes in generated genomes that are highly correlated with published association weights. HELICS provides a scalable generative modeling framework for SNP-level genome data and a path toward conditional synthetic cohorts for data-limited genetic studies. ## RADAgent: Automating RADseq Pipelines with Private Open-Source LLMs DevFeed: [RADAgent: Automating RADseq Pipelines with Private Open-Source LLMs]() Original publisher: [Read original article]() Author: Qiming Guo, Malaka Sudhakara Bava Sukruth Reddy, Honggang Zhao, Ye Wang, Chen Pan, Wenlu Wang Published: 2026-10-27T04:00:00Z Content type: article Language: en Sources: [arXiv: Software Engineering]() Topics: [Data pipelines](), [Large Language Model](), [Open Source](), [AI Integration Strategies]() Tags: [agentic](), [automate](), [control](), [cs-ai](), [cs-se](), [dna](), [execution](), [failure-recovery](), [genome](), [genomics](), [human-oversight](), [llms](), [local](), [modular](), [modules](), [open-source](), [pipelines](), [q-bio-gn](), [quality-control](), [runtime](), [tools](), [validation](), [verification]() ### AI overview RADAgent automates coordination across a complete RADseq analysis using 16 independently executable and verifiable agent modules. Its model-independent runtime bounds tool execution, verifies outputs on disk, and supports failure recovery, addressing workflow state tracking and validation that otherwise require human oversight. Evaluation on a dataset of 476 samples from 19 populations is reported to demonstrate reliable automation with an open-weight LLM, while keeping genomic data and model execution in the local computing environment. ### Source excerpt arXiv:2610.33009v1 Announce Type: cross Abstract: Restriction site-associated DNA sequencing (RADseq) is a family of reduced-representation sequencing methods. It uses restriction enzymes to sample subsets of the genome, therefore allowing multiple individuals to be sequenced and genotyped together at low cost. RADseq is widely used for population genomics, evolutionary biology, marker discovery, and the development of SNP panels, but complete analysis still requires expert coordination across multiple tools and processing stages. Traditional RADseq analysis requires expert coordination of heterogeneous bioinformatics tools, intermediate artifacts, quality-control checks, and parameter-dependent decisions across multiple stages. Although individual steps can be scripted, executing the complete workflow still requires human oversight for state tracking, result validation, and failure handling. To automate this expert-driven coordination, we present RADAgent, which introduces a modular agentic method for reliable RADseq workflow execution with an open-weight LLM. RADAgent maps the natural boundaries of the RADseq workflow into 16 independently executable and verifiable agent modules and uses a model-independent runtime for bounded tool execution, on-disk verification, and failure recovery. We evaluate RADAgent on a real population-genomics dataset containing 476 samples from 19 populations. The results demonstrate that RADAgent reliably automates the complete RADseq workflow with an open-weight LLM while keeping genomic data and model execution within the local computing environment. The code is publicly available at https://github.com/wenlu-lab/RADAgent. ## GenoMAS: Guided Multi-Agent Programming for Scientific Discovery from Genomic Data DevFeed: [GenoMAS: Guided Multi-Agent Programming for Scientific Discovery from Genomic Data]() Original publisher: [Read original article]() Author: Haoyang Liu, Yijiang Li, Haohan Wang Published: 2026-10-21T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Multi Agent Systems](), [AI for science](), [agent orchestration](), [agent to agent communication](), [Data Science]() Tags: [agent](), [cs-ai](), [cs-lg](), [cs-ma](), [data](), [discovery](), [llm](), [multi-agent](), [planning](), [programming](), [q-bio-gn](), [workflows]() ### AI overview GenoMAS combines structured workflows with adaptive LLM agents for genomic analysis. Specialized agents communicate through typed messages, while programming agents break task guidance into Action Units and can advance, revise, bypass, or backtrack as data require. On GenoTEX, it reports 89.13% CSC for preprocessing and 60.48% F1 for trait-associated gene identification, improvements of 10.61% and 16.85% over prior art, respectively. The abstract also reports outperforming strong agent baselines on ScienceAgentBench and DiscoveryBench, and says identified gene-phenotype associations were biologically plausible and supported by literature. ### Source excerpt arXiv:2507.21035v4 Announce Type: replace-cross Abstract: Gene expression analysis holds the key to many biomedical discoveries, yet extracting insights from raw transcriptomic data remains challenging due to the complexity of multiple large, semi-structured files and the need for extensive domain expertise. Current approaches either rely on inflexible workflows that fail in edge cases or fully autonomous agents that lack the precision required for rigorous scientific inquiry. GenoMAS addresses this by presenting a team of LLM-based agents that integrates structured workflows with autonomous adaptability, orchestrating diverse specialized agents through typed message-passing protocols. At its core lies a guided-planning framework where programming agents decompose task guidelines into Action Units and dynamically choose to advance, revise, bypass, or backtrack, maintaining logical coherence while adapting to the idiosyncrasies of genomic data. On the GenoTEX benchmark, GenoMAS achieves 89.13% CSC for data preprocessing and 60.48% $F_1$ for identifying trait-associated genes, surpassing prior art by 10.61% and 16.85%. Beyond transcriptomics, it outperforms strong agent baselines on ScienceAgentBench and DiscoveryBench, two independent bioinformatics benchmarks. Qualitative analysis shows that GenoMAS identifies gene-phenotype associations that are biologically plausible and corroborated by the literature.