benchmark study

Predicted Effector Gene Aggregation, Standards and Unified Schema (PEGASUS): A Community Framework for Effector Gene Reporting

PEGASUS defines a community reporting standard for GWAS predicted effector genes: a metadata record, a full locus-by-gene evidence matrix, and a derived author-prioritised gene list.

Multimodal evidence for impaired glymphatic brain clearance as a mediator of "inflammaging" in schizophrenia - integrating insights from in vivo neuroimaging, omics, and post-mortem single nucleus-RNA-Seq data

The paper integrates neuroimaging, omics, and single nucleus-RNA-Seq data to show glymphatic dysfunction in schizophrenia is linked to inflammation, aging, and cognitive decline.

BatchVaria: variance-based evaluation of batch correction with provenance tracking

BatchVaria evaluates batch correction methods by comparing variance changes across configurations using TCGA breast cancer RNA-seq data.

Rapid whole genome- and annotation-based multilocus sequence typing of the fungal pathogen Histoplasma capsulatum

This paper develops and validates genome-based MLST typing for Histoplasma capsulatum using 406 genome assemblies, achieving lineage separation with >8,000 markers.

GROMTools: scalable individual-level GReX imputation for mega-biobank-scale cohorts

GROMTools, a sparse-weight C++/R engine streaming PLINK2 genotypes, reproduces PrediXcan and PLINK2 GReX imputations with about 100-fold less CPU time and 33-fold less peak memory.

PEtab Select: specification standard and supporting software for automated model selection

PEtab Select provides a standardized format and software for automated model selection in systems biology, supporting diverse modeling frameworks and enabling efficient evaluation of large model spaces.

Platform-specific count-matrix preprocessing workflows for spatial transcriptomics data analysis

This study benchmarks preprocessing workflows for spatial transcriptomics data across multiple platforms and identifies platform-specific recommendations to improve downstream analysis.

Circadian gene-network distortion in high-risk neuroblastoma across multiple biological reference contexts

This study identifies circadian gene network disruption in high-risk neuroblastoma across multiple biological contexts, showing it is a network-level alteration rather than a single-gene effect.

Sum-h2, enabling genetic discovery for deep learning-derived phenotypes through a fast evaluation framework and arena of performance

The paper introduces sum-h2, a fast framework for evaluating genetic discovery from deep learning-derived phenotypes using heritability measures, and establishes a Genetic Discovery Arena for benchmarking.

Inferential boundaries of age prediction: why prediction does not establish biological age measurement

The paper shows that age prediction models do not establish biological age measurement and highlights the need for assumptions beyond age prediction to validate biological age constructs.

Individual Differences in Neural Decoding Are Stable but Instrument Dependent

The paper shows that individual differences in neural decoding stability are influenced by both the person and the measurement instrument, with anatomy and recording characteristics playing key roles.

vep-rs: high-throughput Rust variant annotation with population-scale concordance to Ensembl VEP

vep-rs is a Rust reimplementation of Ensembl VEP with high concordance and significantly improved throughput for variant annotation.

Improving DASH adherence after midlife promotes better late-life cognition and brain health: the SG70 Study

This study shows that better adherence to the DASH diet in midlife and early elderhood improves late-life verbal memory and brain health through structural and neurovascular mechanisms.

PanVasc Research for AI assisted evidence analysis in panvascular intervention

The paper evaluates an AI framework for evidence analysis in panvascular intervention using clinical trial data and NLI4CT tasks, showing improved accuracy with checklist prompts over generic ones.

Where Tabular Foundation Models Falter on Genetic Data: Datasets That Expose and Provide a Path to Address the Gap

The paper evaluates tabular foundation models on genetic data, identifies ancestry-dependent non-stationarity as a key challenge, and proposes synthetic tasks to improve performance across ancestry groups.

AnnoAudit: a marker-based protocol for auditing single-cell atlas annotations reveals systematic, state-dependent annotation failure in a widely used traumatic brain injury resource

A marker audit finds 97.9% of CEREBRI's 45,051 official 'glutamatergic neuron' cells are not glutamatergic (four-check composite 82.6%), creating false ion-channel and OXPHOS signals; true neurons show acute KCNC3 up-regulation.

Accessing Enzyme Kinetic Data and Prediction Methods at Scale

OpenKinetics Predictor unifies thirteen enzyme kinetic predictors behind one interface and API; on E. coli iML1515, eight kcat predictors correlate pairwise at r 0.08 to 0.67.

Serval: A modular framework for decoding imaging based spatial transcriptomics data

Serval is a modular framework for decoding spatial transcriptomics data that introduces Cosine, a method improving transcript recovery and correlation with expression references.

Benchmarking single-cell foundation models for aging biology

This study benchmarks single-cell foundation models for aging biology, finding that Geneformer excels in age prediction and pseudotime concordance, while SCimilarity outperforms others in rare state identification.

RELAX does not reproduce its own estimates at default settings, and its output does not show it

This paper identifies issues with the RELAX software's reproducibility and reliability, showing that it fails to reproduce its own estimates and outputs inconsistent results.

sfate: Schur-free, target-conditioned absorption probabilities for scalable single-cell fate analysis

sfate computes annotation-targeted absorption probabilities by column-wise GMRES on a latent kNN graph, skipping Schur decomposition; a 74,984-cell mouse microglia analysis peaked at 1.76 GiB.

Benchmarking virus identification tools on vaginal metagenomes

This paper benchmarks virus identification tools on vaginal metagenomes, finding geNomad excels in overall classification, VIBRANT minimizes false positives, ViraLM detects short contigs and eukaryotic viruses, and VirSorter2 performs well on VLP-enriched assemblies.

Concordance between the fast and efficient mixed-effects algorithm (FEMA) and conventional mixed-effects modeling for whole-brain connectivity in longitudinal chronic pain

The paper benchmarks the Fast and Efficient Mixed-Effects Algorithm (FEMA) against conventional mixed-effects modeling for longitudinal neuroimaging data and finds high concordance with minor differences in variance component estimates.

Evaluating the reliability of tools for mRNA annotation and IRES studies

A back-splicing circRNA reporter yields linear trans-spliced artifacts; in orthogonal assays three reported cellular IRESes are inactive, and Iso-Seq 5' ends agree with nAnT-iCAGE start sites.

Benchmarking 16S rRNA gene amplicon analysis in high-diversity microbial communities reveals fundamental trade-offs in clustering and denoising

This study benchmarks 16S rRNA gene analysis pipelines in high-diversity datasets, revealing trade-offs between clustering accuracy, species recovery, and abundance reconstruction.

Coevolutionary mining of prokaryotic non-coding elements with a genome language model

Minerva extracts RNA base-pairing, protein-contact and repeat maps from a genome language model without alignments, flagging unannotated structured loci and a phage ncRNA array reverse-transcribed into short cDNAs.

An RNA Language Model trained on sequence alone reveals the structural logic of Internal Ribosome Entry Sites

Continued pretraining of an RNA language model on picornaviral 5′ UTRs alone predicts IRES base pairs with median precision 0.80 versus 0.47 for Infernal, and reveals a Type II subclass.

When Prognostic Compatibility Does Not Guarantee Transfer Utility in Event-Limited Cross-Species Survival Modeling

The paper evaluates cross-species survival modeling in osteosarcoma, finding that prognostic compatibility does not ensure safe transfer utility due to negative transfer and model-selection bias.

Genetic and multi-omic prioritization implicates the CD33-SIGLEC locus in post-traumatic stress disorder

This study identifies CD33-PTSD as a prioritized genetic association through multi-omic analyses and suggests 11 compound candidates for experimental validation.

BOTANIC-1: a series of long-context plant genomic foundation models in the agentic era

Bidirectional Mamba2 plant genomic language models (318M to 3.2B parameters) trained on 320 land-plant genomes top a 22-task frozen-model aggregate; a small variant reaches 131,072 bp context.