benchmark study

Synthetic RNA-seq cohorts for data sharing: a discovery-aware benchmark at transcriptome scale

The paper benchmarks synthetic RNA-seq generation methods across real-world cohorts to evaluate biological fidelity and privacy risks.

Splice-Aware Optimization Prevents Pervasive Missplicing of Natural and Synthetic cDNAs

The paper identifies missplicing as a common issue in heterologous gene expression and develops a splice-aware optimization algorithm to improve transgene performance.

Benchmarking normalisation methods for differential binding analysis in CUT&RUN

This study benchmarks normalization methods for differential binding analysis in CUT&RUN, finding data-driven normalization to be most robust and highlighting variable performance of spike-in normalization.

A Pipeline for Solving Edge-Matching Puzzles and Their Implications for Protein Folding

The paper develops a computational pipeline to solve complex edge-matching puzzles, demonstrating its effectiveness on a puzzle comparable to Eternity II and drawing parallels to protein folding challenges.

Single-cell multi-omic integration analysis prioritizes druggable genes and reveals cell-type-specific causal effects in glioblastomagenesis

The paper integrates GWAS with single-cell multi-omics to identify druggable genes and cell-type-specific causal effects in glioblastoma development.

SpatialCCCbench: Standardized Metrics for the Systematic Evaluation of Spatial Cell-Cell Communication Methods

SpatialCCCbench evaluates spatial cell-cell communication methods using standardized metrics to guide tool selection and benchmark future developments.

Benchmarking the impact of reference genome selection on taxonomic profiling accuracy

This study benchmarks how reference genome selection affects taxonomic profiling accuracy across bacterial and viral datasets, showing context-dependent impacts on accuracy and computational efficiency.

Generalist large language models complement tailor-made predictors for tumor genomics interpretation

This paper evaluates generalist large language models (LLMs) in tumor genomic interpretation tasks, finding they can complement or outperform task-specific models in some cases.

Scaffold-Lab: Critical Evaluation and Ranking of Protein Backbone Generation Methods in A Unified Framework

The paper evaluates and ranks protein backbone generation methods using a unified benchmark, identifying top-performing models in both unconditional and conditional generation.

Selecting methods for draft GEM generation in multicellular eukaryotes: a comparative analysis

This study compares seven automated tools for generating draft genome-scale metabolic models (GEMs) in multicellular eukaryotes, finding no single tool consistently outperforms others.

Benchmarking full-length ITS metabarcoding across Illumina 2x500, PacBio, and Oxford Nanopore sequencing using mock and soil communities

The paper benchmarks full-length ITS metabarcoding across Illumina, PacBio, and Oxford Nanopore platforms using mock and soil communities, finding PacBio has the lowest error rate and recommends standard pipelines for biodiversity assessments.

Sequential acquisition of 1p31.1-p12 LOH and 1q Gain is a common double-hit event in relapsed/refractory myeloma

The paper identifies a therapy-driven genomic double-hit (1pCEN LOH and 1q gain) in relapsed/refractory myeloma, associated with poor prognosis and clonal expansion.

Patient-Calibrated Dynamical Modeling and Embedded Trend-Zone Predictive Control for Closed-Loop Deep Brain Stimulation in Parkinson's Disease

This paper develops a patient-calibrated dynamical model and predictive controller for closed-loop deep brain stimulation in Parkinson's disease, showing improved performance over existing control strategies.

DNAS-Bench: Deterministic Nucleic Acid Screener Benchmarking

DNAS-Bench evaluates Biosecurity Screening Software by testing its ability to detect malicious sequences in synthetic DNA, finding that simple sequence manipulations can bypass many systems.

Bigger Isn't Always Better: Comparing System Size, Hydration, and Software for Lipid Membrane Analysis

This study compares membrane structural properties across different system sizes and hydration levels, and evaluates four analysis software tools for their performance and consistency.

Benchmarking Static Gene Regulatory Network Reconstruction and Dynamic Transition Probing in Single-Cell Foundation Models.

This study benchmarks static gene regulatory network reconstruction and dynamic transition probing in single-cell foundation models across multiple datasets and reference networks.

Microbial network rewiring mediates the function-biosafety trade-off in transboundary Pamir lakes

The paper examines microbial network rewiring in transboundary Pamir lakes, revealing trade-offs between ecosystem function and biosafety under environmental pressures.

Calibrated Prediction Intervals for Polygenic Scores: Updated Comparisons, Contextual Calibration, and Data Normalization

This paper compares two methods for creating calibrated prediction intervals for polygenic scores, showing their performance and the importance of normalization and preprocessing.

Benchmarking strain-level profiling of Escherichia coli in short-read gut metagenomes

This study benchmarks six tools for strain-level profiling of E. coli in short-read gut metagenomes, showing significant variation in performance and highlighting StrainR2's high accuracy and PathoScope 2.0/StrainGE's robustness in unknown strain scenarios.

Building an Interoperable Rare Disease Multi-omic Resource: The GREGoR Data Model and Dataset

The paper introduces a data model and dataset for rare disease research, enabling interoperable multi-omic data integration across clinical sites.

BiomniBench: Process-level Evaluation of LLM Agents for Real-world Biomedical Research

BiomniBench evaluates LLM agents' research processes in biomedical tasks using expert rubrics, revealing performance gaps in method selection and reasoning.

Integrative Genomic Analyses Identify COL21A1 and ENPEP-FGF5 Regulatory Pathways for Blood Pressure Variation in East Asians

Cosine Similarity Conflates Clinically Distinct Cancer Variants: A Case for Typed-Graph Retrieval in Precision Oncology Decision Support

This paper shows cosine similarity in RAG systems conflates clinically distinct cancer variants, proposing typed-graph retrieval as a more accurate alternative for precision oncology.

MetaNet: a scalable and integrated tool for reproducible omics network analysis

MetaNet is an R package for scalable multi-omics network analysis, offering faster computation and better memory efficiency than existing tools, with applications in microbial and exposome-transcriptome studies.

A unified benchmark of synthetic data generation for clinical transcriptomic cancer cohorts

This paper benchmarks synthetic data generation methods for clinical transcriptomic cancer data, finding no single method outperforms all metrics, with Gaussian Copula showing the most balanced performance.

Widespread DNA off-targeting confounds studies of RNA chromatin occupancy

The paper shows that many reported RNA-chromatin interactions are technical artifacts caused by DNA off-targeting, not true biological interactions.

Accelerated neural aging as a systems-level vulnerability for tinnitus

This study identifies accelerated neural aging as a risk factor for tinnitus by analyzing MEG and structural imaging data from multiple datasets.

Benchmarking siRNA Prediction: The Role of Representation and Validation Strategies

The paper benchmarks siRNA prediction methods, showing that rigorous cross-validation and z-curve representation improve performance and reliability.

The impact of long-read sequencing on fungal genome assemblies: progress and disparity

This paper reviews the impact of long-read sequencing on fungal genome assemblies, highlighting progress, disparities, and remaining taxonomic gaps.

Benchmarking Boltz-2 for Screening of Therapeutic Antibody-Antigen Interactions

The paper benchmarks Boltz-2 for screening therapeutic antibody-antigen interactions, finding weak but above-chance discrimination and limited benefit from extensive sampling.