AI 中文总结
研究旨在解决基因检测变异解释难及G2P门户可视化不便问题。核心方法是构建原生iOS框架G2P Explorer，它能在设备上解析相关数据并通过多模块呈现结果。主要贡献是提供多种功能，支持离线且性能良好，成为G2P门户的移动伴侣。
AI 中文摘要
基因检测报告编码变异的速度远远快于其可被解释的速度，将变异置于其生物物理背景中，如它干扰的结构域、残基是埋藏还是暴露、是否靠近二硫键或预测的结合口袋等，越来越需要将其投影到三维蛋白质模型上。基因组学2蛋白质（G2P）门户将基因到结构标识符链与一个密集的、按残基索引的注释表统一起来，但其可视化层假定为桌面浏览器，在床边、教室或野外使用不便，因为手机或平板电脑通常是唯一可用设备。我们展示了G2P Explorer，这是一个原生iOS框架，它在设备上使用公共G2P REST API，无损解析其71列制表符分隔的特征表，并通过共享单个可观察视图模型的六个相互关联的模块呈现结果。除了移植，它还贡献了一个SwiftUI Canvas多轨道序列渲染器、门户不可用的异构体比对路线的设备上重建、吸收AlphaFold模型文件版本控制方案的双向Swift JavaScript结构桥以及解析半结构化自由文本并区分缺失注释和零注释的容错摄取层。每个搜索的蛋白质在首次获取后缓存在设备上，因此可以立即重新打开并离线工作，嵌入式结构视图以适合小屏幕的简化形式绘制。在跨越189 - 1863个残基的六种蛋白质上，该框架维持交互式帧时间（3.1 - 16.6毫秒）和适度内存（16 - 72MB）。G2P Explorer是G2P门户的一个开放、可重现的移动伴侣，用于假设生成、教学和随时随地的变异解释。
英文摘要
Genetic testing reports coding variants far faster than they can be interpreted, and placing a variant in its biophysical context, the domain it perturbs, whether its residue is buried or exposed, whether it lies near a disulfide bond or a predicted binding pocket increasingly requires projecting it onto a three dimensional protein model. The Genomics 2 Proteins (G2P) portal unifies the gene-to-structure identifier chain with a dense, residue indexed annotation table, but its visualization layer presumes a desktop browser and is awkward at the bedside, in the classroom, or in the field, where a phone or tablet is often the only device. We present G2P Explorer, a native iOS framework that consumes the public G2P REST API on device, parses its seventy one column tab separated feature tables without loss of fidelity, and presents the result through six interlinked modules sharing a single observable view model. Beyond porting, it contributes a SwiftUI Canvas multi-track sequence renderer, an on device reconstruction of the portal's unavailable isoform alignment route, a bidirectional Swift JavaScript structural bridge that absorbs the AlphaFold model file versioning scheme, and a fault tolerant ingestion layer that parses semi structured free text and distinguishes absent annotations from zero. Each searched protein is cached on device after the first fetch, so it reopens instantly and works offline, and the embedded structural view is drawn in a reduced form suited to a small screen. Across six proteins spanning 189-1{,}863 residues, the framework sustains interactive frame times (3.1-16.6\,ms) and modest memory (16-72\,MB). G2P Explorer is an open, reproducible mobile companion to the G2P portal for hypothesis generation, teaching, and on the go variant interpretation.
