Custom Microbial Reference Database and Predicted peptides across 11cancer types (part 1) for "Microbial transcriptional reprogramming defines tumour-restricted antigens from taxonomically overlapping microbiomes"

This repository supports the study titled "Microbial transcriptional reprogramming defines tumour-restricted antigens from taxonomically overlapping microbiomes" by providing: Antigen Analysis Results: Microbial antigen predictions across 11 cancer types (including in-house CRC and pan-cancer datasets). The file Pan-CancerMicrobialAntigensAnalysis-MHCII_part2.zip is available at [Zenodo link: 10.5281/zenodo.15043056]. Custom Microbial Reference Datasets: A curated database of 4,342 species (545 genera, ~10M proteins), filtered for human-tumor relevance using LLM-based annotation and manual curation. Includes microbial genomes and protein reference sequences.本仓库为研究论文《微生物转录重编程定义来自分类学上重叠的微生物组的肿瘤限制性抗原》提供支持，具体内容如下： 抗原分析结果： 针对11种癌症类型（包含自研结直肠癌（Colorectal Cancer，CRC）及泛癌症数据集）开展的微生物抗原预测分析。 文件Pan-CancerMicrobialAntigensAnalysis-MHCII_part2.zip 可于[Zenodo链接：10.5281/zenodo.15043056]获取。 定制化微生物参考数据集： 经整理的4342个物种（545个属，约1000万条蛋白质）数据库，通过基于大语言模型（Large Language Model）的注释与人工整理筛选，遴选出与人类肿瘤相关的菌株。 该数据库包含微生物基因组与蛋白质参考序列。