Phage Functional Protein Candidate Database v1.0

Phage Functional Protein Candidate Database v1.0 Abstract / 摘要 A prediction-level candidate database of phage functional proteins, covering three therapeutically relevant classes: depolymerases, endolysins, and holins. It contains 4,117 candidate entries, each ranked by a six-dimension scoring scheme (novelty, transfer/translational value, and overall total score). For every candidate the database provides genomic context, functional-annotation evidence (Pfam / CDD / superfamily / CAZy), structural evidence (homology fold, mean pLDDT, predicted ligand pockets), and ML-assisted inference fields, so that users can triage high-value candidates for wet-lab validation. 本库为预测级候选库(in silico predicted),收录噬菌体功能蛋白候选 4117 条,聚焦解聚酶、内溶素类、穿孔素三类,按六维评分排序。每条候选均提供基因组背景、功能注释证据、结构证据与机器学习辅助推断字段,用于湿实验前的优先级筛选。 Data files / 数据文件 · 候选数据库_v1.0.csv — 4,117 data rows (+ header), UTF-8. · 候选数据库_v1.0.sqlite — SQLite database, table `candidates` (4,117 rows). · README.md — schema and reproduction queries. Both files carry identical records; the CSV is for spreadsheet users, the SQLite for query/reproducibility. Field schema / 字段口径 Field Meaning 含义 category Functional class: depolymerase / endolysin / holin 功能类别 host_raw / host_domain / genome_family Host name, host domain (e.g. bacteria), and genome family 宿主与分类 best_evalue Best E-value from homology search 最优同源 E 值 n_pfam / n_cdd / n_superfamily Counts of Pfam / CDD / SUPERFAMILY hits 功能注释证据数 cazy_layer CAZy carbohydrate-active enzyme layer annotation CAZy 糖酶层注释 tmhmm_predhel_approx Approximate transmembrane helix count (heuristic, marked approx) 跨膜螺旋数(启发式估计) signal_peptide_approx Signal peptide prediction (heuristic, marked approx) 信号肽预测(启发式估计) fold_homolog Structural fold homology annotation 结构折叠同源注释 mean_plddt Mean pLDDT from structure prediction 结构预测平均置信度 n_pockets Number of predicted ligand-binding pockets 预测结合口袋数 substrate_inferred Inferred substrate (in silico inferred, not experimentally confirmed) 底物推断(计算推断,非实验确定) novelty / transfer_value / total_score Three of the six scoring dimensions: novelty, translational value, overall total 新颖性 / 转化价值 / 总分 Important caveats / 重要说明 · The database is prediction-level (in silico predicted); substrate assignments are inferred, not experimentally validated. · Topology/secretion annotations (tmhmm_predhel_approx, signal_peptide_approx) are heuristic approximations. · Score dimensions are relative ranking aids, not absolute activity measures. Reproducibility / 复现查询(SQLite) -- Top-100 high-value candidates (clinical priority hosts)SELECT protein_id, category, host_raw, total_score FROM candidatesWHERE host_domain='bacteria' ORDER BY total_score DESC LIMIT 100;-- High-novelty depolymerases (no strong homolog)SELECT protein_id, novelty, fold_homolog FROM candidatesWHERE category='depolymerase' AND novelty>=0.9; How to cite / 引用方式 Please cite this dataset as: LPS318. (2026). Phage Functional Protein Candidate Database v1.0 [Data set]. Zenodo. https://doi.org/10.5281/zenodo.21971299 · Cite the DOI above (10.5281/zenodo.21971299) in your manuscript or supplementary materials. · If future versions (v1.1.0, …) are released, please cite the Concept DOI (the persistent DOI that always points to the latest version). · Source code/repository: https://github.com/LPS318/phage-functional-proteins