benchmark study

Fundamental Limits of Inferring Dynamical Gene Regulatory Models from Single-Cell Data

This study evaluates methods for inferring Jacobian matrices from single-cell data to model gene regulatory dynamics, finding that while some methods capture behavioral properties, they fail to reliably capture structural properties.

Benchmarking of Optimization Algorithms for Boolean Model Inference in Biomedicine

This paper benchmarks four Boolean network optimization methods for gene regulatory modeling, showing distinct performance in accuracy, efficiency, and robustness to perturbations.

DELPHAI, AI Agent for Predicting Drug Response and Resistance

DELPHAI predicts drug response and resistance using single-cell perturbation data from organoids, identifying mechanisms and recommending therapies.

Understanding Foundation Models in Digital Pathology: Performance, Trade-offs, and Model-Selection Recommendations

This study benchmarks five digital pathology foundation models across various tasks, showing that model size does not always correlate with performance, and recommends selecting models based on task-specific needs.

A globally influential area-condition metric is a poor proxy for invertebrate biodiversity

The paper evaluates a biodiversity metric used for ecological compensation and finds it poorly correlates with invertebrate biodiversity metrics like richness and abundance.

Pipeline Olympics: continuable benchmarking of computational workflows for DNA methylation sequencing data against an experimental gold-standard

The paper benchmarks DNA methylation sequencing workflows against an experimental gold-standard, identifying superior workflows and developing a sustainable benchmarking platform.

Hill-type models of skeletal muscle and neuromuscular actuators: a systematic review

This paper reviews Hill-type muscle models, finding most lack innovation, proper validation, parameter identification, and reusability.

Benchmarking long-read RNA-sequencing technologies with LongBench: a cross-platform reference dataset profiling cancer cell lines with bulk and single-cell approaches

LongBench benchmarks long-read RNA-sequencing technologies using a cross-platform dataset of cancer cell lines, showing high gene-level concordance but reduced transcript-level consistency across platforms.

Splicing neoepitope prediction is sensitive to methodological differences

The paper evaluates the impact of methodological differences on splicing neoepitope prediction, finding high variability and limited reproducibility across pipelines.

Performance evaluation of adaptive introgression classification methods

This study evaluates adaptive introgression classification methods using simulated datasets from different evolutionary scenarios, finding that Q95-based methods are most efficient for exploratory analysis.

A Benchmark of Evo2 Genomic AI Models for Efficient and Practical Deployment

This paper benchmarks the Evo2 genomic AI model in cancer genomics, showing FP8 precision on H800 GPUs improves inference speed and accuracy, while larger models perform worse on non-FP8 hardware.

Lost in a Large EEG Multiverse? Comparing Sampling Approaches for Representative Pipeline Selection

The paper compares sampling methods for selecting representative pipelines in multiverse analyses of EEG data to assess their impact on result robustness.

Development of an Integrated Single-Cell and Spatial Transcriptomics Atlas of Healthy Human Skin Focusing on the Pilosebaceous Unit

The paper develops an integrated single-cell and spatial transcriptomics atlas of healthy human skin, focusing on the pilosebaceous unit, and identifies previously under-characterized cell types and signaling hubs.

AI mirrors experimental science to uncover a novel mechanism of gene transfer crucial to bacterial evolution

This paper tests an AI's ability to generate hypotheses about bacterial gene transfer mechanisms, finding it matches experimentally confirmed results.

Comparative Analysis of Single-Nucleus and Single-Cell RNA Sequencing in Human Bone Marrow Mononuclear Cells: Methodological Insights and Trade-offs

This study compares single-nucleus and single-cell RNA sequencing in human bone marrow mononuclear cells, identifying method-specific biases and trade-offs.

Availability of benign missense variant truthsets for validation of functional assays: current status and a novel systematic approach

The paper evaluates the limitations of existing benign missense variant truthsets for MAVEs and proposes a proactive-systematic framework to improve their construction for better functional evidence classification.

Diversity and abundance of ring nucleases in type III CRISPR-Cas loci

The paper analyzes type III CRISPR-Cas loci across 38,742 prokaryotic genomes to identify and characterize ring nucleases that regulate cyclic oligoadenylate levels.

Functional relevance of CASP16 nucleic acid predictions as evaluated by structure providers

Structure providers for 18 CASP16 nucleic acid targets find blind predictions capture secondary structure but miss non-canonical interactions, backbone kinks, ligand sites, and intermolecular interfaces.

"I'd like to think I'd be able to spot one if I saw one": How science journalists navigate predatory journals

This study examines how science journalists identify and avoid predatory journals through interviews, finding they rely on journal prestige and quality rather than awareness of predatory practices.

Creating a biomedical knowledge base by addressing GPT inaccurate responses and benchmarking context

GNQA, a retrieval-augmented GPT question-answering system over 3000 GeneNetwork, aging and diabetes papers, cites source paragraphs; RAGAS rated answer relevance above human thumbs-up ratings (89% vs 76%).

BioML-bench: Evaluation of AI Agents for End-to-End Biomedical ML

BioML-bench evaluates AI agents on end-to-end biomedical ML tasks across four domains, finding agents underperform compared to humans and diverse ML strategies correlate with higher scores.

Genome-wide pleiotropy analysis of longitudinal blood pressure and harmonized cognitive performance measures

This paper identifies genome-wide pleiotropic loci linking blood pressure and cognitive performance measures using longitudinal data from multiple cohorts.

The current landscape and emerging challenges of benchmarking single-cell methods

This paper benchmarks single-cell methods by analyzing 282 papers, highlighting challenges in combining benchmarking studies and preventing fatigue.

tMHG-Finder: Tree-guided Maximal Homologous Group Finder for Bacterial Genomes

tMHG-Finder uses Mash guide trees and consensus representative sequences to find maximal homologous groups in bacterial genomes with linear scaling, reaching F-score 0.932 on simulated data.

Multi-omics time-series analysis in microbiome research: a systematic review

This paper reviews data science methods for longitudinal multi-omics in microbiome research, highlighting the need for integrated approaches and the lack of widely adopted methods.

Benchmarking large language models for genomic knowledge with GeneTuring

This paper benchmarks large language models for genomic knowledge using GeneTuring, finding that a custom GPT-4o setup with NCBI APIs performs best.

What Large Language Models Know About Plant Molecular Biology

This paper benchmarks large language models in plant molecular biology using a new dataset, finding strong performance on multiple-choice questions but limitations in factual accuracy and self-awareness.

Benchmarking of proximity-dependent biotinylation enzymes across cellular compartments and time windows

This study benchmarks ten proximity ligases across cellular compartments and time windows, finding variations in activity, background, and spatial specificity, with ultraID showing the best balance.

IDEAL-GENOM: Integrated Downstream Analytical Toolkit for Genomic Analysis

IDEAL-GENOM streamlines GWAS analysis by integrating tools like PLINK and GCTA into a Python-based framework for reproducible and customizable genomic analysis.

Signatures of Electron-hole Hopping in Myoglobin Peroxidase Activity Revealed by Deep Mutational Learning

This study uses deep mutational learning to identify myoglobin variants with enhanced peroxidase activity, showing the role of electron-hole hopping in catalytic efficiency.