Non-Redundant Cryo-EM Test Dataset (2026 release targets)
Year: 2026
Type: dataset
Abstract: This dataset consists of a curated collection of 12 non-redundant structural targets released in 2026, specifically assembled to benchmark 3D structure prediction, alignment, and macromolecular modeling tools. Dataset Composition & Filtering Criteria: Release Date: Restricted strictly to structures released on or after January 1, 2026. Experimental Method: Electron Microscopy (EM). Resolution Limits: Map resolutions bounded between $1.8\,\text{\AA}$ and $3.8\,\text{\AA}$ to ensure sufficient density detail for automated sequence assignment. Polymer Type: Pure protein assemblies (Protein only). Monomeric Length: Total chain lengths bounded between 400 and 1,800 amino acid residues. Redundancy Control: Clustered and verified at a strict 25% sequence identity threshold using MMseqs2 to ensure independent structural topologies. Contained Files (Benchmark_2026_Test_Dataset.zip): Experimental reference coordinate models (.cif) Reference FASTA sequences (.fasta) Primary 3D cryo-EM density maps (.map) Target metadata summary tables including recommended map contour levels Reproduction & Executable Notebook: The complete benchmarking workflow, pipeline setup, statistical tables, and performance evaluations across all six modeling methods (EModelX, MICA, ModelAngelo, CryoAtom, EMProt, and CryoZeta) are documented and executable via the companion GitHub Jupyter notebook: Cryo-EM Density Map Benchmark Notebook. Reproduction & Executable Notebook:The complete benchmarking workflow, pipeline setup, statistical tables, and performance evaluations are documented and executable via the companion GitHub Jupyter notebook: https://github.com/adelekeadeniyan/cryo-em-benchmarking-2026/blob/main/docs/Cryo-EM%20Density%20Map%20Benchmark.ipynb.
Author: Adeniyan, Adeleke
Institution: Wesley University
Language: English
Subfield: Structural Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7222817420