DIA-LiPQuan ACS supplementary-data workbench demo
DIA-LiPQuan ACS 补充数据工作台 Demo

中文
1. 定量值来自 ACS 论文 Table S1/S2 的 DIA-NN DDALib peptide 工作表；
   报告中的线性强度为原表 log2 强度的 2^x 转换，缺失值保持缺失。
2. 复合 FASTA 按论文方法构建：UP000005640（人）和 UP000002311（酵母）的
   reviewed 条目，再拼接 cRAP v1.0 实验室类 1 和灰尘/接触类 2 污染物。
   本地 UniProt 2026_02 提供 20,329 条人源、
   6,066 条酵母及 50/57 条 cRAP；
   与论文 2024 年 10 月版本允许有轻微序列/条目版本差异，明细见 composite-fasta-provenance.tsv。
   cRAP accession 使用 CRAP1_/CRAP2_ 命名空间避免覆盖物种蛋白；目标 FASTA 不含 decoy，
   decoy 应由上游 FragPipe、DIA-NN 或 Spectronaut 搜索步骤生成。
3. 打开 https://gemma.omicsolution.com/dia-lipquan.html，选择“载入 Demo 输入并练习提交”，
   可先查看与真实上传一致的字段识别、样品映射和质量检查。
4. 实际上传时选择 report 和 FASTA；experiment-layout 可以一并上传，也可以省略后
   从未压缩 report TSV 的 Run 列在线识别样品并编辑条件和重复。
5. 实验设计包含 H0、H2、H5、H7.5、H30 各 4 个样品运行；人源补充表没有 H0，
   因此 H0 仅保留酵母来源条目，不会人为补造人源定量值。
6. 默认按 H2_vs_H5、H7.5_vs_H5、H30_vs_H5 的论文顺序分析。论文式汇总可合并展示，
   火山图始终一次只显示一个掺比，并通过图内按钮切换。
7. 选择 DIA-NN、|log2FC|=1、组内校正 P=0.05，完成“上传并检查输入”后提交，
   可获得与即时 Demo 相同的 Figure 2、Figure 3A、差异汇总、CV、五类结果表和交互图表。
8. 本输入包先按蛋白进行确定性簇抽样，再按物种、肽段类型和三个比较的可定量缺失模式校准配额；
   抽样哈希不使用任何 FC、P 值或差异结论，避免用待评价结果反向选择肽段。
9. 这是便于在线练习的代表性子集；完整论文图表复核数据另行提供。

English
1. Quantities come from the DIA-NN DDALib peptide sheets in ACS Tables S1/S2.
   Linear report intensity equals 2^x of the source log2 intensity; missing values remain missing.
2. The composite FASTA follows the paper: reviewed entries from UP000005640 (human)
   and UP000002311 (yeast), plus cRAP v1.0 laboratory type 1 and dust/contact type 2 contaminants.
   Local UniProt 2026_02 contributes 20,329 human,
   6,066 yeast, and 50/57 cRAP entries.
   Minor sequence or entry-version differences from the paper's October 2024 release are accepted;
   see composite-fasta-provenance.tsv. cRAP accessions are namespaced as CRAP1_/CRAP2_ to prevent
   collisions. No decoys are included; the upstream search engine should generate them.
3. Open https://gemma.omicsolution.com/dia-lipquan.html and select Load demo inputs
   and practice submission to preview the same field, sample, and quality checks used after upload.
4. For a manual upload, select the report and FASTA. The experiment layout may be uploaded,
   or omitted and generated online from the Run column of the uncompressed report TSV.
5. The design contains four runs each for H0, H2, H5, H7.5, and H30. The human
   supplementary table has no H0 values, so H0 retains yeast entries only; no human values are fabricated.
6. H2_vs_H5, H7.5_vs_H5, and H30_vs_H5 are analyzed in paper order. Paper-style summaries
   can combine ratios, while the volcano plot always shows one ratio at a time with in-chart switching.
7. Select DIA-NN, |log2FC|=1, and group-wise adjusted P=0.05. After Upload and validate
   inputs, submit to obtain the same Figure 2, Figure 3A, differential, CV, five-table, and interactive views.
8. This practice package first selects deterministic protein clusters, then calibrates quotas
   by species, peptide type, and three-comparison quantifiability pattern. Hash selection never uses FC, P values,
   or differential outcomes, avoiding outcome-informed peptide selection.
9. This is a representative web-sized practice subset; the complete paper-figure validation
   package is provided separately.
