遇见数据集

ApexOracle paper assets: embeddings, MIC scoring, and result-reproduction data

收藏
Zenodo2026-08-11 更新2026-08-13 收录
官方服务:

资源简介:

This version extends the ApexOracle paper-asset series while retaining all files from earlier versions. Genome_embs.tar.gz and Text_Description.tar.gz: paper-listed precomputed pathogen assets. apexoracle_fig1b_classification_reproduction_v1.tar.gz: exact Fig. 1b folds, predictions and metric checker. apexoracle_fixed_mic_reconstruction_v1.tar.gz: post-paper fixed-split MIC reconstruction. apexoracle_synergy_replay_v1.tar.gz: complete 3-fold by 7-member synergy checkpoint replay with privacy-minimized predictions and an independent metric checker. The synergy seed-0 split reproduces every archived fold metric after rounding to four decimals and is released as a high-confidence historical candidate, not as proven exact 2025 membership. Machine-readable file roles, hashes and scientific boundaries are provided in the version release manifest. Version 5.0.0 adds source-partitioned public model-ready MIC, small-molecule classification, and synergy tables, plus the compact strain mapping and paper genome list. The private in-house MIC partition is excluded.

提供机构:
Zenodo
创建时间:
2026-08-11
二维码
社区交流群
二维码
科研交流群
商业服务