Bioinformatics Advances

Motivation: Protein--protein interactions (PPIs) underpin a wide range of biological functions in living organisms. Experimental identification of new PPIs is expensive and time-consuming. The experimental bottleneck has implied an imbalance in terms of data availability: while certain species have been screened exhaustively, other species have not been sufficiently examined. An AI driven protocol for PPI prediction that leverages the massive data accumulated for certain species means a decisive boost for so far understudied species. Results: We present ANIMA (Artificial Neural Interaction Model for Animals), an AI supported cross-species PPI prediction model trained on popular species to predict PPIs in under-researched species. Our experiments demonstrate that our model, when trained on 200 diversely selected animal species, can successfully predict PPIs in other species: ANIMA achieves 95.3% accuracy on other animal species, 91.1% on other eukaryotes, and 82.5% on non-eukaryotes. For a more fine-grained evaluation of the model, we stratify performance rates by the evolutionary distance of test to training sets. We also stratify results by a novel, alignment based score ("representation score") which allows for fine-grained evaluation in terms of its capacity to generalize to unseen interactions. As expected, results demonstrate increasing performance on increasing evolutionary similarity and on increasing identity of interacting proteins, while still showing excellent performance on proteins entirely lacking counterparts in the training set. In comparison with the state of the art, ANIMA demonstrates substantial superiority in terms of performance rates.

Background: Over-representation analysis (ORA) is widely used to interpret gene lists from high-throughput biological experiments. However, ORA often produces long and fragmented lists of enriched terms that are difficult to translate into coherent, testable biological hypotheses. This challenge is amplified by the rapid growth of gene set collections and the need to integrate results across multiple gene-set collections. Large language models (LLMs) can assist in summarizing enrichment outputs, but their effectiveness is limited by the lack of reproducibility and statistically supported representations. Results: We introduce iterative Gene Enrichment Analysis (iGEA), a software framework that transforms ORA-based enrichment outputs into structured, interpretable networks, supporting human interpretation and AI-assisted exploration. iGEA iteratively selects the most significant enriched term, removes its overlapping genes from the input list, and repeats enrichment until no significant terms remain. Applied independently across gene-set collections, this procedure yields a compact set of non-overlapping enriched terms within each collection. Integration of collection-specific results yields a cross-collection gene-term network in which genes connect terms from different collections. Using a published set of HIV dependency factors, iGEA identified five compact modules spanning secretory trafficking, nuclear transport, transcription elongation, proteostasis, and innate immune signaling, enabling rapid hypothesis generation and interactive exploration. The resulting network structure supports standardized prompting and provides a structured representation for LLM-assisted summarization and exploration of enrichment results. Conclusions: iGEA provides a software framework for gene enrichment analysis that addresses the interpretation challenge of ORA by generating compact, empirically benchmarked cross-collection gene-term networks. This network representation reduces within-collection redundancy, exposes relationships across gene-set collections, and provides a structure that makes modules and hub genes easier to identify through visualization, network-based analysis, and LLM-assisted exploration. Availability: Source code is available at: https://github.com/aion-labs/Gene-Enrichment-Analysis A web-based version of the application is available at: https://iterative-gene-enrichment.streamlit.app

Biomedical knowledge graphs can connect information across genes, phenotypes, tissues, pathways, experiments, and clinical resources, but they are difficult to query correctly without detailed knowledge of the graph. Large language models can help write Cypher code for knowledge graphs, yet a query focused on a bioinformatics task that looks reasonable may still use the wrong identifier, relationship direction, source, intermediate node, or output unit if the user is not completely trained on the schema. We developed ddkg.skill, an Agent Skill that can be loaded into a compatible LLM session to support querying of the NIH Common Fund Data Ecosystem Data Distillery Knowledge Graph (DDKG), a biomedical property graph integrating more than 180 ontologies, over 40 genomics datasets, and dozens of additional cross-domain biomedical datasets. Rather than utilizing a simple markdown prompt, ddkg.skill contains a controller, a compositional library of DDKG-specific references and structured tables, primary documentation, validated query patterns, a routing table, and a Python script that checks the internal links among these materials. The skill build for the December 2025 DDKG release contains 38 bundled files and 239 routing relationships and is identified by checksum so users can state the exact build used to compose a query. The evaluation reported here was performed separately on an earlier build with 38 files and 227 routing entries. In that orthogonal nine-test evaluation against a live December 2025 DDKG instance, five of seven tests targeting sources absent from the worked examples produced correct executed results. The evaluation also exposed a failed query, a cross-source comparison that was not biologically well posed, and errors in the skills own reference material that informed later revisions. The skill guides an AI through entity resolution, graph inspection, query construction, and validation for biomedical and bioinformatics queries, while keeping the DDKG itself as the source of returned results. This design provides a portable and versioned method for giving general-purpose AI systems practical knowledge of a complex biomedical knowledge graph to empower complex bioinformatics data integration. Three executed biomedical use cases further demonstrate cross-species phenotype-to-expression querying, genomic overlap between 4DN chromatin-loop anchors and GTEx eQTLs, and integration of congenital heart phenotype associations with ClinVar and GTEx heart evidence.

Sequence similarity networks (SSNs) are graphical representations of sequence relationship frequently used for exploring protein sequence space. Conventional SSN workflows typically use BLAST to calculate sequence similarities and rely on external visualization tools to gener-ate the final networks. Consequently, raw sequence data is often detached from the calculat-ed similarities during visualization, complicating SSN analyses that require residue-level in-formation. To bridge this gap, we present EMAP-SSN, an open-source, cross-platform soft-ware suite that integrates SSN computation, visualization, and analyses in streamlined work-flows. The program provides BLAST- and embedding-based alignment pipelines for se-quence-similarity calculation and directly links network nodes to their original sequences and multiple alignments for analyses. Modular architectures for embedding generation, com-mand integration, and browser-based utilities allow additions of research-specific functionalities and facilitate future development. Using a set of fold-type IV pyridoxal 5'-phosphate-dependent enzymes, we demonstrate how EMAP-SSN connects network topology with residue-level variation to identify sequence clusters, map functional motifs, and detect subgroup-specific conservation patterns. These capabilities provide a practical route from large protein sequence sets to experimentally verifiable hypotheses on enzyme function and targets for protein engineering. The EMAP-SSN program can be accessed from https://github.com/Xuebin-Feng/EMAP-SSN.

Small RNA (sRNA) sequencing is widely used to study siRNA, miRNA and piRNA pathways and to profile sRNAs derived from viruses, transposons and host transcripts. However, many sRNA workflows rely on command-line mapping and separate software for downstream analysis and visualisation. We present sRNAMap, an HTML5/JavaScript application for mapping and analysing sRNA sequencing data. sRNAMap accepts either raw FASTA/FASTQ reads or pre-aligned BAM files. For raw reads, sRNAMap uses a k-mer-seeded alignment algorithm that runs directly in the browser using parallel Web Workers where supported. It generates strand-specific per-base coverage, 5'- and 3';-end profiles, read-length distributions and nucleotide-bias summaries, with optional library-size normalisation. It also quantifies characteristic read-overlap signatures, including 5';-to-5'; overlaps associated with ping-pong-like processing. Additional modules quantify phased sRNA production, non-templated 3'- and 5'-additions, and sequence diversity. Analyses can be saved and restored as JSON or exported as self-contained HTML reports containing figures and analysis metadata. Benchmarking against Bowtie 2 under approximately matched alignment settings using a representative virus-infected arthropod small-RNA dataset showed highly concordant mapping profiles. Per-position depth showed an R^2 of approximately 0.998, while 5';- and 3';-end counts showed R^2 values of approximately 0.97. sRNAMap is available at https://github.com/rhparry/sRNAMap under the MIT licence.

Metastasis involves context-dependent molecular interactions in which non-coding RNAs, particularly miRNAs and circRNAs, play important regulatory roles. However, existing computational approaches generally do not jointly represent cancer type, metastatic event, and cancer-specific metastatic context. We developed a context-aware multi-task heterogeneous graph neural network (GNN) for predicting ncRNA associations with cancer types and metastatic events. The framework integrates multiple biological repositories into a heterogeneous graph representing ncRNAs, cancers, metastatic event types (METs), and cancer-specific metastatic instances (CSMIs). The model performs six link-prediction tasks using a hierarchical transformer-based encoder and multi-relational TuckER decoder. Across ten independently initialized runs evaluated on the RNA-group-disjoint held-out test set, the model achieved a global AUROC of 0.8801 {+/-} 0.0118 and an F1 score of 0.8260 {+/-} 0.0071. All three ablation variants yielded lower AUROC, with the largest reduction under independent task training. Case studies in pancreatic cancer, colorectal cancer, and hepatocellular carcinoma provided disease-level, event-level, and expression-based support, respectively, for top-ranked candidate associations. The framework enables context-specific prioritization of ncRNA-cancer-metastasis associations for experimental evaluation.

Motivation: LLM agents increasingly draft functional interpretations of gene lists, but can cite Gene Ontology (GO) terms that no current enrichment backend returned for that list, and can pair real GO accessions with fabricated labels. Results: We present GroundAnnot, a client for PANTHER, Enrichr, and g:Profiler that returns a closed vocabulary of GO term IDs and their backend labels, and enforces two contracts: Contract A (no ID or label outside the backend payload) and Contract B (no enrichment claim outside the backend's significant set). Across three open-weight models (Qwen2.5-7B local; Qwen3.8-27B and GPT-OSS-120B via Groq) on six curated disease gene lists, mean valid-enriched rates were 15.3%, 20.1%, and 35.1%. All unsupported IDs were real GO terms classified by QuickGO as wrong-biology, obsolete, or wrong-branch (zero fabricated accessions). A distinct failure mode emerged: for the Alzheimer's gene list, the local 7B model paired every one of its ten stable picks with a label that does not match the current GO term for that accession (10/10 mismatches), producing a coherent synaptic-signalling narrative that the IDs do not support. Mean overlap with the backends' FDR top-10 union was 0.33-0.67/10. On 50 MSigDB Hallmark gene sets, the strongest model named the defining pathway in 21/44 (47.7%) directly named cases. When an enriched shortlist was placed in the prompt, all three models complied fully (60/60 for each model; 180/180 overall), showing that a bounded vocabulary removes these output classes when the downstream agent is required to use it.

Agentic artificial intelligence (AAI) is increasingly used by biomedical scientists, where it has substantially lowered the difficulty of integrating computational, statistical and data-science approaches into day-to-day research activities, assisting with project design, analysis and manuscript preparation. Significant barriers remain, however, which are unlikely to be removed solely by releasing more capable AI models. Meaningful use still demands a technical fluency that many biomedical researchers lack: a PI and their lab members must learn how to use the AAI system, establish standard operating procedures (SOPs) for managing, sharing and curating data, and enforce rules that address data privacy and security. Existing commercial AAI systems, which were designed primarily for software engineering purposes, need to be significantly re-purposed for biomedical-related data science. We present Murmurent, shared software that sits beneath the agentic AI and provides the skills, procedures and protections to address the concerns above, removing the need for a lab to develop them independently. This includes (1) infrastructure to support projects and so-called choreographies} involving multiple group members; (2) a set of specialized agents each dedicated to typical biomedical data science tasks including software and statistical development, data visualization, literature search, equity, diversity, inclusion, and decolonization (EDID) review, and others; (3) a tiered memory which can retain important research decisions, intermediate, derived outputs and sensitive data, and which can exploit this information to automatically build better contexts in AAI sessions; (4) traceability records which are used by the AAI session to better plan and execute data analyses and software builds by, for example, helping to avoid repeating decisions that lead to dead-ends; (5) enforcement of SOPs for data maintenance and data governance across all lab members; and (6) multi-user capacity to allow groups to interact and collaborate. We use the system to identify putative inhibitors of Peptidyl-prolyl cis-trans Isomerase NIMA-interacting 1 (Pin1), a protein with a shallow catalytic site making it difficult to target. We describe the construction of several approaches to identify Pin1 inhibitors and the results they yield. Murmurent is open source.

Graph embedding transforms complex networks into low-dimensional representations for analysis and downstream machine learning. Existing methods often rely on eigenvalue ordering, random-walk sampling, or neural optimization. We introduce MOSAIC, Multistep Operator Spectral Accumulation with Isolated Components, an unsupervised framework that constructs an accumulated diffusion operator from successive powers of a graph transition matrix. MOSAIC uses Krylov subspace analysis to identify dominant spectral components according to their convergence behavior. This approach emphasizes persistent and spectrally separated diffusion modes rather than selecting embedding coordinates solely by eigenvalue magnitude. We evaluated MOSAIC using a biomedical graph projection derived from the Data Distillery Knowledge Graph that contained 5,171 human genes and 10,147 phenotypes. Across embedding dimensions from 16 to 128, MOSAIC achieved node-type classification AUROCs from 0.9863 to 0.9926 and exceeded FastRP, Node2Vec, HashGNN, and GraphSAGE. For gene-phenotype link prediction, validation and test associations were withheld before embedding generation. Under progressive removal of all edges or specific edge classes, MOSAIC retained more node-label information than FastRP, including after extreme graph sparsification. In a focused DICER1 analysis, 106 of 137 documented phenotype associations received scores above 0.90, demonstrating concordance with known pleiotropic biology and prioritizing additional relationships for independent evaluation. These results support MOSAIC as a fully unsupervised framework for learning informative graph representations. Although demonstrated using a biomedical graph projection, MOSAIC is domain independent and can be applied to other complex networks

AO_SCPLOWBSTRACTC_SCPLOWMicroRNAs (miRNAs) and small interfering RNAs (siRNAs) share Argonaute-mediated guide-target recognition, yet quantitative siRNA efficacy measurements are substantially scarcer and more costly to generate than miRNA-target interaction data. We therefore asked whether miRNA interaction data could provide transferable supervision for siRNA efficacy prediction. Here we present DuplexFM, a biologically grounded framework that uses sample-specific gates to integrate five evidence sources: pairing and sequence-context priors, duplex energetics, experimentally supervised mRNA accessibility, target-to-guide cross-attention, and contextual token-pair compatibility. The accessibility expert, trained on nucleotide-resolution icSHAPE measurements, achieved a held-out nucleotide-level Pearson correlation of 0.627 and evaluated accessibility at seed match and energy-supported candidate sites. On miRBench v7, three independently trained DuplexFM models achieved a macro APS of 0.873{+/-}0.002, soft-voting increased this to 0.876 and yielded the highest APS on all four test sets. We then froze the miRNA-trained representation and trained only a lightweight residual head with 24 siRNA-specific descriptors. Transfer improved Pearson and Spearman correlations, AUPRC, and F1 over the descriptor-only baseline in all six evaluation settings. The ensemble achieved the highest Pearson and Spearman correlations in four settings, whereas OligoFormer remained stronger on Huesken and Takayuki. These findings show that experimentally grounded accessibility and miRNA-derived interaction representations provide complementary, transferable information, supporting a parameter-efficient route towards unified modeling of Argonaute-guided RNA regulation. Code and data are available at https://github.com/cbaiming/DuplexFM.

Biopharmaceutical companies routinely use forward screening to identify potential ligands for targets of biological consequence. Increasingly, they are also using reverse screening to proactively identify downstream off-target liabilities and new repurposing opportunities. Exhaustive reverse docking of one or a multitude of molecules across every characterized site is one approach, but the computational cost is often prohibitive. One molecule against 28,579 receptor sites takes 40.5 hours on twenty CPU cores. We present a method that puts a retrieval step in front of the docking. Every co-crystallized ligand in the Protein Data Bank is indexed by the 3-dimensional pharmacophore fingerprint it presents, together with the UniProt accessions it was solved against, and a query retrieves the sites belonging to its nearest neighbors. Fingerprints are predicted from two-dimensional structures with PharmCast, so a query is fingerprinted in 4 ms and searched against 27,797 indexed ligands in 40 ms. For a query molecule we take its 5 most similar indexed ligands and pool every protein those 5 were crystallized with, which averages 21.1 distinct proteins. Across 3,000 held-out molecules that pool contained the molecule's own known target 48.8 percent of the time. Docking those 21.1 proteins takes 108 s, against 40.5 hours for the full panel. Pooling the twenty-five most similar ligands instead gives 104.5 proteins and finds the target 60.8 percent of the time. To be indexed, a structure only has to establish which ligand sat in which protein. Holding the index to pocket-grade coordinates had been excluding whole receptor classes. Admitting X-ray at 2.5 [A] and cryo-EM at 4.0 [A] takes coverage from 3,670 target sites to 28,579. Two 2025 clinical molecules were cross-validated. Orforglipron returned the glucagon-like peptide 1 (GLP-1) receptor first of 27,797 ligands, through a non-identical analog at 0.901; daraxonrasib returned the KRAS and cyclophilin A tri-complex fourth, at 0.836. When run in batches the pipeline reverse screens about 96,000 molecules per hour on 1 core, roughly 2.3 million per day, so the retrieval step is practical even with a massive virtual library. The method is available at reversescreen.ai, which allows users the opportunity to screen molecules against the current index, returns the retrieved sites, and docks them individually. The index itself can be downloaded for confidential screening in a local environment. Keywords: reverse screening, target identification, pharmacophore fingerprint, molecular docking, polypharmacology, off-target prediction, virtual screening, repurposing

Motivation: Graph Neural Networks (GNNs) have gained increasing interest in the biomedical domain, as the integration of prior knowledge and deep neural networks has the potential to enhance insights into molecular processes and disease mechanisms. However, a comprehensive and systematic assessment of model architectures, data modalities, graph structures, and their performance for graph signal classification in the biomedical domain is yet to be performed. In order to close this gap, we conducted a benchmarking study on multiple GNNs on a Protein-Protein Interaction (PPI) network for Kidney Renal Clear Cell Carcinoma and Breast cancer subtype prediction, performing an in-depth investigation of architectures, incorporating skip connections and various data modalities. Results: While none of the GNNs outperforms the structure-agnostic Multi-Layer Perceptron baseline, all of them can handle bimodal data (gene methylation and expression) and offer the ability to gain explainability based on PPIs. We offer practical guidelines for applying GNNs to graph signal processing tasks specifically for cancer classification. Depending on the underlying dataset and PPI structure employed, models on different data modalities outperform others. Overall, we suggest using ChebNet, which tends to outperform the Graph Convolutional Network and the Graph Attention Network in cancer subtype prediction. We recommend using GNN architectures that employ a simple flattening readout layer, as they provide better classification performance and faster training time than those with global average pooling. Additionally, we tested residual connections, but they had only an insignificant impact on classification performance.

Drug discovery depends on coordinated decisions across target validation, structure analysis, molecular design, developability assessment and synthetic feasibility, but current computational methods often operate as disconnected tools. Here, we introduce TRACEDD (Tool-grounded Reasoning and Agentic Coordination for Explainable Drug Design), a framework that makes three primary contributions: (1) It establishes a 'tool-first' multi agentic architecture where LLMs orchestrate validated computational tools rather than replace them, ensuring scientific rigor. (2) It implements a multi-agent system that mirrors expert discovery teams, enabling transparent and traceable decision-making through a Reason-Act-Observe loop. (3) It demonstrates an end-to-end workflow, from target validation to synthesis planning, that adaptively handles real-world data variability, such as the absence of experimental structures. The framework decomposes discovery into specialized agents for target validation, druggability assessment, molecular generation, lead optimization, ADMET evaluation, literature evidence integration and retrosynthesis, all operating through a Reason Act Observe workflow. Using JAK2 as a representative case, we show that the system can retrieve experimental protein structures, invoke AlphaFold when structures are unavailable, identify druggable pockets and perform de novo molecular generation. Known JAK2 inhibitors are used to define design hypotheses and guide reinforcement learning-based molecular generation, with docking scores/predicted pIC50 and other physicochemical/ADMET properties serving as reward and prioritization signals. The framework demonstrates a tool-first, reasoning-driven approach in which each major decision is linked to explicit tool invocation, intermediate evidence. By combining agentic orchestration with domain-specific computational tools, the system supports transparent, adaptable and human-verifiable molecular design workflows, providing a foundation for more reliable AI-assisted drug discovery.

Interpreting a bacterial methylome is a multi-omics problem. It requires integrating modified-base calls with genome annotation, motif inventories, methyltransferase genotypes, transcript abundance, replichore position and, increasingly, chromosome conformation. These data types are commonly generated in incompatible formats, use inconsistent sequence and gene identifiers, and originate from different analytical workflows. The relevant algorithms are available, but assembling them into a coherent and statistically defensible analysis remains a substantial data-integration and interface problem. We present MErlin (Methylation-driven Expression & Regulation Linkage in Interacting Nuclear-domains), a multi-omics toolkit comprising seventeen composable modules, from basecalled modBAM files to ranked gene-level evidence and a self-contained HTML report. MErlin is distributed both as a conventional Python package and as an agent skill: a structured, version-controlled layer of procedural knowledge that enables a compatible large language model (LLM) assistant to select and operate the audited package without generating a new analysis implementation for each request. This design treats natural language as an interface to fixed analytical operations rather than as a substitute for tested scientific software. The skill encodes module-selection rules, mandatory preflight checks, questions that require human input, design-to-inference constraints, and interpretation guidance. We describe MErlin's architecture and statistics, validate it against a synthetic dataset with planted ground truth, and illustrate the conversational interface on a real methylome-transcriptome comparison in Pseudoalteromonas haloplanktis TAC125. MErlin is open source and available at https://github.com/IacopoPasseri/MErlin

Network-based analyses of molecular interactions are useful for interpreting high-throughput omics data and identifying therapeutic targets. Cytoscape is the standard platform for these tasks, but users face a trade-off between accessible graphical workflows that are difficult to document and reproducible automation in Python or R that requires programming expertise. General-purpose coding assistants can generate Cytoscape Automation scripts, but remain external to Cytoscape. We present CyChat, a Cytoscape Desktop app that integrates a chat interface and a large language model (LLM) agent into the application. CyChat translates natural language into executable Cytoscape Automation workflows, runs generated Python code, and exports chat sessions with executed code as standalone Jupyter notebooks. To reduce setup barriers, CyChat includes an embedded Python runtime and supports both cloud-based and locally hosted LLMs. CyChat was evaluated across ten Cytoscape workflows using seven LLM providers, each represented by one LLM. The strongest configuration achieves a pass rate above 99%. In a qualitative evaluation based on a published network visualization, CyChat completes the task in 1.5-5 minutes, compared with 15-20 minutes for manual GUI workflows by computational biologists. CyChat is available through the Cytoscape App Store at https://apps.cytoscape.org/apps/cychat.

AlphaFold3 has shown promise as a tool for predicting antibody-antigen binding, yet its performance across large datasets has not been fully characterized. In this study, 3401 experimentally validated antibody-antigen complexes were sourced from the Structural Antibody Database and screened alongside 23798 negative controls to benchmark AlphaFold3s binding prediction capabilities. Confidence metrics including Predicted Aligned Error and Interface Predicted Template Modeling score were used to achieving a maximum recall of 53% at 100 inference seeds. Several factors were found to influence prediction accuracy: a notable bias was observed toward antibodies derived from X-ray crystallography structures versus those from electron microscopy, and positive prediction rates were found to decrease with increasing target protein size and surface area. In contrast, neither the amino acid composition or lengths of the complementarity determining regions, nor training data leakage were found to introduce significant bias. An innate false positive rate of approximately 3% was identified, with AF3 shown to hallucinate plausible binding interfaces across the surface of decoy targets while avoiding disordered regions. Epitope mapping using DockQ, epitope shift, and antibody displacement revealed that approximately 34% of false negatives retained the correct epitope location despite poor structural alignment, suggesting that conformation refinement tools could recover additional true binding predictions. These findings provide a comprehensive characterization of AlphaFold3s strengths and limitations for antibody screening in computational drug discovery. Key MessagesO_LIAlphaFold3 has a recall of 50% and an innate false positive prediction rate of 3%. C_LIO_LIFalse negative predictions can still feature the correct epitope despite poor RMSD. C_LIO_LIFactors such as disorder and target size impact accuracy. C_LI

Single-cell RNA sequencing has enabled the construction of comprehensive cell atlases, yet the quality and coherence of the cell-type annotations within these atlases remain largely unexamined. When a label is applied to a transcriptionally heterogeneous population, the downstream analyses that depend on it, and automated label transfer in particular, become unreliable. We present scINTILLA (Single-Cell Integrated Inference, Labelling, and Landscape Analysis), a computational framework that combines supervised and unsupervised machine learning to score the learnability and internal consistency of cell-type labels in single-cell datasets. The unsupervised arm benchmarks a broad panel of clustering algorithms and derives a neighbourhood confusion score for every cell, whilst the supervised arm trains up to twelve classifiers and extracts prediction agreement, entropy, and confidence. These signals are normalised and aggregated into a single composite score per cell type, where a low score flags label ambiguity or concealed heterogeneity. As a by-product, scIN-TILLA also reports which clustering and classification algorithms perform best on a given dataset, offering practical guidance for downstream label transfer. We applied it to five Human Cell Atlas datasets spanning the adult brain, lung, eye, and two organoid atlases, and recovered clear differences in the learnability and internal consistency of annotations across atlases that were not driven by the number of annotated cell types. Focused re-analysis of lowscoring populations in the lung and endoderm-organoid atlases resolved biologically coherent sub-populations, in some cases with context-specific enrichment, much of it recovered from cells that had been assigned broad or catch-all labels. scINTILLA is advisory rather than prescriptive, guiding principled, data-driven re-annotation at atlas scale.

Summary: Automated protein function annotation remains challenging as sequence databases outpace curated labels and homology-based transfer fails for proteins lacking close relatives. We present GOlien, a scalable, composition-based method that annotates protein sequences using a Shannon-entropy k-mer model. Built from the CAFA3-based training split and evaluated on the held-out validation split, GOlien achieves micro-averaged of 0.7317 (Biological Process), 0.7610 (Cellular Component) and 0.8295 (Molecular Function). Our tool offers a practical, complementary alternative to existing pipelines and extending annotation coverage for proteins. Availability and Implementation: The command line tool to submit FASTA files and retrieve predicted GO term annotations with the corresponding documentation is available at: https://github.com/probalytiq/golien-tool

Predicting cellular responses to genetic perturbations is central to understanding gene function and prioritizing therapeutic targets, but experimental screens cannot exhaustively cover genes, cell types, and perturbation combinations. Recent benchmarks have shown that simple baselines can match or outperform substantially more complex models, suggesting that informative biological priors may be as important as model capacity. Here we present SLIM, a lightweight extension of the bilinear model of Ahlmann-Eltze et al. SLIM represents perturbations with 64-dimensional embeddings derived from the STRING protein network and predicts mean transcriptional responses through a closed-form ridge-regression estimator. It then constructs single-cell populations by retrieving training cells and rescaling each gene to match the predicted mean. We evaluated SLIM against four deep learning models and two simple baselines on four single-gene perturbation datasets and one combinatorial perturbation dataset. Across these within-dataset benchmarks, SLIM achieved competitive mean-response accuracy, ranked first in eight of twelve single-gene dataset-metric comparisons, and produced substantially lower maximum mean discrepancy values than the evaluated alternatives. The model has 640 trainable parameters and fitted each benchmark dataset in under 10 seconds on a CPU. These results show that compact biological representations can support accurate and computationally efficient perturbation prediction. Code is available at https://github.com/RasmussenLab/SLIM. Key PointsO_LISLIM combines a closed-form bilinear predictor with STRING-derived perturbation embeddings. C_LIO_LIAcross five within-dataset benchmarks, SLIM achieved competitive mean-response prediction with only 640 trainable parameters. C_LIO_LISLIM builds cell populations by rescaling retrieved training cells to the predicted mean, so they inherit realistic cell-to-cell variation and gene-gene covariation. C_LIO_LIThe results highlight the importance of perturbation representations and population-construction procedures in low-data benchmarks. C_LIO_LISLIM fits each benchmark dataset in under 10 seconds on a standard CPU. C_LI

Recent advances in artificial intelligence have enabled models to capture protein sequence and structural features. We present PeTriPPI2, a hybrid framework for protein--protein interaction (PPI) prediction that combines sequence embeddings from ESM-2 with structural representations from PeTriBERT, an encoder originally designed for inverse folding. We evaluate this sequence--structure approach on the Pinder dataset. On a test set of 2,342 protein pairs, the selected checkpoint achieves 0.898 accuracy, 0.917 precision, 0.876 recall, and 0.896 F1 score for the interacting class (AUROC 0.963). Compared with SpatialPPIv2 on the same test data, PeTriPPI2 has higher precision (0.917 versus 0.886) but lower recall (0.876 versus 0.942) and F1 score (0.896 versus 0.913). Component ablations indicate that both the sequence and structural pathways contribute to this configuration.