AI x Bio Discovery
@aixbiobot.bsky.social
Automated discovery of AI x Bio preprint papers.
Genome-wide single-cell perturbation screens with VIPerturb-seq [new]
...are now routine through probe-based detection and phenotypic enrichment, using combinatorial indexing for expanded throughput.
...are now routine through probe-based detection and phenotypic enrichment, using combinatorial indexing for expanded throughput.
February 14, 2026 at 11:56 AM
Genome-wide single-cell perturbation screens with VIPerturb-seq [new]
...are now routine through probe-based detection and phenotypic enrichment, using combinatorial indexing for expanded throughput.
...are now routine through probe-based detection and phenotypic enrichment, using combinatorial indexing for expanded throughput.
Unifying Non-Equilibrium Information Thermodynamics and Genome Engine Dynamics: Maxwell's Demon Control of Cancer Cell Fates [new]
Maxwell's Demon crit. pt., orders cancer genome dynamics for global coherence & irreversible fate.
Maxwell's Demon crit. pt., orders cancer genome dynamics for global coherence & irreversible fate.
February 14, 2026 at 11:54 AM
Unifying Non-Equilibrium Information Thermodynamics and Genome Engine Dynamics: Maxwell's Demon Control of Cancer Cell Fates [new]
Maxwell's Demon crit. pt., orders cancer genome dynamics for global coherence & irreversible fate.
Maxwell's Demon crit. pt., orders cancer genome dynamics for global coherence & irreversible fate.
CodonRL: Multi-Objective Codon Sequence Optimization Using Demonstration-Guided Reinforcement Learning [new]
...learns struct. priors from folding feedback & expert sequences, enabling user-ctrl tradeoffs for enhanced mRNA design.
...learns struct. priors from folding feedback & expert sequences, enabling user-ctrl tradeoffs for enhanced mRNA design.
February 14, 2026 at 11:30 AM
CodonRL: Multi-Objective Codon Sequence Optimization Using Demonstration-Guided Reinforcement Learning [new]
...learns struct. priors from folding feedback & expert sequences, enabling user-ctrl tradeoffs for enhanced mRNA design.
...learns struct. priors from folding feedback & expert sequences, enabling user-ctrl tradeoffs for enhanced mRNA design.
Discovery of TDP-43 aggregation inhibitors via a hybrid machine learning framework [new]
integrating GNN, descriptors, & annotations to identify berberrubine & PE859, inhibiting aggregation in cells and C. elegans.
integrating GNN, descriptors, & annotations to identify berberrubine & PE859, inhibiting aggregation in cells and C. elegans.
February 14, 2026 at 11:08 AM
Discovery of TDP-43 aggregation inhibitors via a hybrid machine learning framework [new]
integrating GNN, descriptors, & annotations to identify berberrubine & PE859, inhibiting aggregation in cells and C. elegans.
integrating GNN, descriptors, & annotations to identify berberrubine & PE859, inhibiting aggregation in cells and C. elegans.
Feature-based in-silico model to predict the Mycobacterium tuberculosis bedaquiline phenotype associated with Rv0678 variants [new]
Uses 5 feats (evo conserv, funct site prox) to interpret Rv0678 var impact & guide clin mgmt.
Uses 5 feats (evo conserv, funct site prox) to interpret Rv0678 var impact & guide clin mgmt.
February 14, 2026 at 11:06 AM
Feature-based in-silico model to predict the Mycobacterium tuberculosis bedaquiline phenotype associated with Rv0678 variants [new]
Uses 5 feats (evo conserv, funct site prox) to interpret Rv0678 var impact & guide clin mgmt.
Uses 5 feats (evo conserv, funct site prox) to interpret Rv0678 var impact & guide clin mgmt.
February 14, 2026 at 11:04 AM
Democratised single cell proteomics uncovers tumor associated macrophage signature in CYLD cutaneous syndrome skin tumors [new]
Accessible, cost-effective, label-free tech found novel TME remodeling macrophage pop in CCS skin tumors.
Accessible, cost-effective, label-free tech found novel TME remodeling macrophage pop in CCS skin tumors.
February 14, 2026 at 11:03 AM
Democratised single cell proteomics uncovers tumor associated macrophage signature in CYLD cutaneous syndrome skin tumors [new]
Accessible, cost-effective, label-free tech found novel TME remodeling macrophage pop in CCS skin tumors.
Accessible, cost-effective, label-free tech found novel TME remodeling macrophage pop in CCS skin tumors.
Evolution of oncogene amplification across 86,000 cancer cell genomes [new]
...shows distinct evo paths for ICamp & ecDNA amplifications: ICamp has subclonal spec.; ecDNA has variable, rearrangement-driven, tissue-spec. evolution.
...shows distinct evo paths for ICamp & ecDNA amplifications: ICamp has subclonal spec.; ecDNA has variable, rearrangement-driven, tissue-spec. evolution.
February 14, 2026 at 10:40 AM
Evolution of oncogene amplification across 86,000 cancer cell genomes [new]
...shows distinct evo paths for ICamp & ecDNA amplifications: ICamp has subclonal spec.; ecDNA has variable, rearrangement-driven, tissue-spec. evolution.
...shows distinct evo paths for ICamp & ecDNA amplifications: ICamp has subclonal spec.; ecDNA has variable, rearrangement-driven, tissue-spec. evolution.
Cell phenotypes in the biomedical literature: a systematic analysis and text mining corpus [new]
Annotated corpus of literature cell phenotypes links them to Cell Ontology, revealing naming patterns and enabling ontology expansion.
Annotated corpus of literature cell phenotypes links them to Cell Ontology, revealing naming patterns and enabling ontology expansion.
February 14, 2026 at 9:14 AM
Cell phenotypes in the biomedical literature: a systematic analysis and text mining corpus [new]
Annotated corpus of literature cell phenotypes links them to Cell Ontology, revealing naming patterns and enabling ontology expansion.
Annotated corpus of literature cell phenotypes links them to Cell Ontology, revealing naming patterns and enabling ontology expansion.
ChatDIA: A zero-shot large language model workflow for targeted analysis of data-independent acquisition mass spectrometry data [new]
automates targeted DIA peptide ID & interactive data expl. via zero-shot LLM reason chromatograms.
automates targeted DIA peptide ID & interactive data expl. via zero-shot LLM reason chromatograms.
February 14, 2026 at 9:01 AM
ChatDIA: A zero-shot large language model workflow for targeted analysis of data-independent acquisition mass spectrometry data [new]
automates targeted DIA peptide ID & interactive data expl. via zero-shot LLM reason chromatograms.
automates targeted DIA peptide ID & interactive data expl. via zero-shot LLM reason chromatograms.
A single-cell atlas and aging clock define biological age and risk-associated stem cell states in human hematopoiesis [new]
HSPC atlas to define age programs. An ML aging clock predicts age; AML dev. links to gen. risk & survival.
HSPC atlas to define age programs. An ML aging clock predicts age; AML dev. links to gen. risk & survival.
February 14, 2026 at 8:59 AM
A single-cell atlas and aging clock define biological age and risk-associated stem cell states in human hematopoiesis [new]
HSPC atlas to define age programs. An ML aging clock predicts age; AML dev. links to gen. risk & survival.
HSPC atlas to define age programs. An ML aging clock predicts age; AML dev. links to gen. risk & survival.
TOXsiRNA: A web server to predict the toxicity of chemically modified siRNAs [new]
by computationally predicting their toxicity and off-targets using machine learning models trained on various siRNA sequence features.
by computationally predicting their toxicity and off-targets using machine learning models trained on various siRNA sequence features.
February 14, 2026 at 8:57 AM
TOXsiRNA: A web server to predict the toxicity of chemically modified siRNAs [new]
by computationally predicting their toxicity and off-targets using machine learning models trained on various siRNA sequence features.
by computationally predicting their toxicity and off-targets using machine learning models trained on various siRNA sequence features.
evoCancerGPT: Generating Zero-Shot Single-Cell and Single-Sample Cancer Progression Through Transfer Learning [new]
...by forecasting gene exp profiles from pseudotime-ordered sc states to charact. tumor prog. in indiv. pts.
...by forecasting gene exp profiles from pseudotime-ordered sc states to charact. tumor prog. in indiv. pts.
February 14, 2026 at 8:56 AM
evoCancerGPT: Generating Zero-Shot Single-Cell and Single-Sample Cancer Progression Through Transfer Learning [new]
...by forecasting gene exp profiles from pseudotime-ordered sc states to charact. tumor prog. in indiv. pts.
...by forecasting gene exp profiles from pseudotime-ordered sc states to charact. tumor prog. in indiv. pts.
MassID provides near complete annotation of metabolomics data with identification probabilities [new]
cloudbased untargeted end-to-end pipeline using deep learning & DecoID2 for FDR-controlled probabilistic metabolite ID.
cloudbased untargeted end-to-end pipeline using deep learning & DecoID2 for FDR-controlled probabilistic metabolite ID.
February 14, 2026 at 8:53 AM
MassID provides near complete annotation of metabolomics data with identification probabilities [new]
cloudbased untargeted end-to-end pipeline using deep learning & DecoID2 for FDR-controlled probabilistic metabolite ID.
cloudbased untargeted end-to-end pipeline using deep learning & DecoID2 for FDR-controlled probabilistic metabolite ID.
Machine learning-guided design of artificial microRNAs for targeted gene silencing [new]
developing a web platform using NNs and large-scale data to optimize amiRNA sequence-structure-processing for precise, specific gene silencing.
developing a web platform using NNs and large-scale data to optimize amiRNA sequence-structure-processing for precise, specific gene silencing.
February 14, 2026 at 8:51 AM
Machine learning-guided design of artificial microRNAs for targeted gene silencing [new]
developing a web platform using NNs and large-scale data to optimize amiRNA sequence-structure-processing for precise, specific gene silencing.
developing a web platform using NNs and large-scale data to optimize amiRNA sequence-structure-processing for precise, specific gene silencing.
Enzyme Classification via Semi-Supervised Functional ResidueLearning [new]
...by learning sparse functional residue activations via MSA-based data augmentation for robust EC prediction and interpretable annotations.
...by learning sparse functional residue activations via MSA-based data augmentation for robust EC prediction and interpretable annotations.
February 14, 2026 at 8:49 AM
Enzyme Classification via Semi-Supervised Functional ResidueLearning [new]
...by learning sparse functional residue activations via MSA-based data augmentation for robust EC prediction and interpretable annotations.
...by learning sparse functional residue activations via MSA-based data augmentation for robust EC prediction and interpretable annotations.
Parameter-free representations outperform single-cell foundation models on downstream benchmarks [new]
...via simple linear models on norm. scRNA-seq to capture cell ID across diverse/novel tasks, challenging complex deep learning.
...via simple linear models on norm. scRNA-seq to capture cell ID across diverse/novel tasks, challenging complex deep learning.
February 14, 2026 at 8:26 AM
Parameter-free representations outperform single-cell foundation models on downstream benchmarks [new]
...via simple linear models on norm. scRNA-seq to capture cell ID across diverse/novel tasks, challenging complex deep learning.
...via simple linear models on norm. scRNA-seq to capture cell ID across diverse/novel tasks, challenging complex deep learning.
Pragmatic analysis with knowledge-guided for unraveling peptide-protein pairwise non-covalent mechanisms [new]
...integrates contextual information and biological prior knowledge to build residue-level interaction maps.
...integrates contextual information and biological prior knowledge to build residue-level interaction maps.
February 14, 2026 at 8:13 AM
Pragmatic analysis with knowledge-guided for unraveling peptide-protein pairwise non-covalent mechanisms [new]
...integrates contextual information and biological prior knowledge to build residue-level interaction maps.
...integrates contextual information and biological prior knowledge to build residue-level interaction maps.
Enhancing Pan-cancer Spatial Transcriptomics atSingle-cell Resolution with stPainter [new]
..., achieved via a pre-trained gen model that reconstructs genome-wide expression from sparse spatial data, enabling univ zero-shot analysis.
..., achieved via a pre-trained gen model that reconstructs genome-wide expression from sparse spatial data, enabling univ zero-shot analysis.
February 14, 2026 at 8:11 AM
Enhancing Pan-cancer Spatial Transcriptomics atSingle-cell Resolution with stPainter [new]
..., achieved via a pre-trained gen model that reconstructs genome-wide expression from sparse spatial data, enabling univ zero-shot analysis.
..., achieved via a pre-trained gen model that reconstructs genome-wide expression from sparse spatial data, enabling univ zero-shot analysis.
ADAMIXTURE: Adaptive First-Order Optimization for Biobank-Scale Genetic Clustering [new]
employs EM with adaptive first-order gradients, approximating curvature for efficient biobank-scale genetic clustering.
employs EM with adaptive first-order gradients, approximating curvature for efficient biobank-scale genetic clustering.
February 14, 2026 at 8:09 AM
ADAMIXTURE: Adaptive First-Order Optimization for Biobank-Scale Genetic Clustering [new]
employs EM with adaptive first-order gradients, approximating curvature for efficient biobank-scale genetic clustering.
employs EM with adaptive first-order gradients, approximating curvature for efficient biobank-scale genetic clustering.
PathoResist AI: A One-Click Web Platform for Rapid Pathogen Resistance Analysis Based on the all_ratio Algorithm [new]
...analyzes nanopore seq data w/ all-ratio algorithm for automated pathogen ID & AMR gene pred. frm FASTQ uploads.
...analyzes nanopore seq data w/ all-ratio algorithm for automated pathogen ID & AMR gene pred. frm FASTQ uploads.
February 14, 2026 at 8:08 AM
PathoResist AI: A One-Click Web Platform for Rapid Pathogen Resistance Analysis Based on the all_ratio Algorithm [new]
...analyzes nanopore seq data w/ all-ratio algorithm for automated pathogen ID & AMR gene pred. frm FASTQ uploads.
...analyzes nanopore seq data w/ all-ratio algorithm for automated pathogen ID & AMR gene pred. frm FASTQ uploads.
Disentangled multimodal evolutionary representations for cross-virus predictive modeling of antigenic change [new]
learned via flow-based gen. framework integrating seq. homology, physchem. & struct. features predict immune escape.
learned via flow-based gen. framework integrating seq. homology, physchem. & struct. features predict immune escape.
February 14, 2026 at 8:06 AM
Disentangled multimodal evolutionary representations for cross-virus predictive modeling of antigenic change [new]
learned via flow-based gen. framework integrating seq. homology, physchem. & struct. features predict immune escape.
learned via flow-based gen. framework integrating seq. homology, physchem. & struct. features predict immune escape.
ARGformer: learning on ancestral recombination graphs with transformers [new]
...by creating contextual embeddings from ARGs for population genetic analysis, enabling unsupervised ancestry inference & archaic introgression detection.
...by creating contextual embeddings from ARGs for population genetic analysis, enabling unsupervised ancestry inference & archaic introgression detection.
February 14, 2026 at 7:42 AM
ARGformer: learning on ancestral recombination graphs with transformers [new]
...by creating contextual embeddings from ARGs for population genetic analysis, enabling unsupervised ancestry inference & archaic introgression detection.
...by creating contextual embeddings from ARGs for population genetic analysis, enabling unsupervised ancestry inference & archaic introgression detection.
Benchmarking Machine Learning and Automated Image Analysis for Organelle Quantification [new]
...reveals pipeline variability and morph/modality dep., offering practical guidance for robust, reproducible organelle imaging workflows.
...reveals pipeline variability and morph/modality dep., offering practical guidance for robust, reproducible organelle imaging workflows.
February 14, 2026 at 7:41 AM
Benchmarking Machine Learning and Automated Image Analysis for Organelle Quantification [new]
...reveals pipeline variability and morph/modality dep., offering practical guidance for robust, reproducible organelle imaging workflows.
...reveals pipeline variability and morph/modality dep., offering practical guidance for robust, reproducible organelle imaging workflows.
Disulphide and sequence-encoded conformational priors guide nanobody structure prediction [new]
...by incorporating predictable HCDR3 conformational "blueprints" and non-canonical disulphide bonds as explicit modeling constraints.
...by incorporating predictable HCDR3 conformational "blueprints" and non-canonical disulphide bonds as explicit modeling constraints.
February 14, 2026 at 7:18 AM
Disulphide and sequence-encoded conformational priors guide nanobody structure prediction [new]
...by incorporating predictable HCDR3 conformational "blueprints" and non-canonical disulphide bonds as explicit modeling constraints.
...by incorporating predictable HCDR3 conformational "blueprints" and non-canonical disulphide bonds as explicit modeling constraints.