Rename GENOS-EVEE to GENOS-VarRisk

modules/pipeline/EXTERNAL_PATHS.md

@@ -75,7 +75,7 @@ cp .env.example .env
 | 04 | `04_vep/raw_vep.tsv` | 原始 VEP TSV |
 | 04 | `04_vep/vep.log` | VEP 日志 |
 | 05 | `05_vcf_info_to_csv/vep_output.with_info.csv` | 合并 VCF INFO/FORMAT |
-| 06 | `06_genos_evee_annotation/vep_output.with_genos_evee.csv` | GENOS-EVEE 宽表 |
+| 06 | `06_genos_evee_annotation/vep_output.with_genos_evee.csv` | GENOS-VarRisk 宽表 |
 | 07 | `07_hla_filter/vep_output.no_hla.csv` | **最终输出**（默认 `hla_filter=yes`） |
 | 汇总 | `full_pipeline.outputs.tsv` | 各步路径索引 |
 | 日志 | `logs/full_pipeline.log` | 全流程日志 |

modules/pipeline/README.md

@@ -1,6 +1,6 @@
 # OpenRare Pipeline — 罕见病变异注释与宽表输出全流程
 
-从患者 VCF 出发，串联 Beagle phasing、VCF 前置处理、假基因注释、VEP 多插件注释、INFO 回填、GENOS-EVEE 注释与 HLA 过滤，一键产出可解读的宽表 CSV；支持命令行与 FastAPI 两种调用方式。
+从患者 VCF 出发，串联 Beagle phasing、VCF 前置处理、假基因注释、VEP 多插件注释、INFO 回填、GENOS-VarRisk 注释与 HLA 过滤，一键产出可解读的宽表 CSV；支持命令行与 FastAPI 两种调用方式。
 
 ## 目录
 
@@ -123,7 +123,7 @@ full_log                      /path/to/output/logs/full_pipeline.log
 最终 CSV 表头示例（列较多，此处仅示意）：
 
 ```text
-#CHROM,POS,REF,ALT,...,Gene,Consequence,CLIN_SIG,...,GENOS-EVEE,...
+#CHROM,POS,REF,ALT,...,Gene,Consequence,CLIN_SIG,...,GENOS-VarRisk,...
 chr1,1197557,G,A,...,TTLL10,missense_variant,...,0.42,...
 ```
 
@@ -315,7 +315,7 @@ pixi run bash scripts/run_full_pipeline.sh --help
 | 03 | `03_pseudogene_annotation/` | 假基因 INFO 注释（可跳过） | `preprocessed.pseudogene_annotated.vcf.gz` |
 | 04 | `04_vep/` | VEP + CADD/SpliceAI/dbNSFP/LoFTEE/ClinVar 等 | `vep_output.base.csv` |
 | 05 | `05_vcf_info_to_csv/` | VCF INFO/FORMAT 回填到 CSV | `vep_output.with_info.csv` |
-| 06 | `06_genos_evee_annotation/` | GENOS-EVEE 疾病预测分数 | `vep_output.with_genos_evee.csv` |
+| 06 | `06_genos_evee_annotation/` | GENOS-VarRisk 疾病预测分数 | `vep_output.with_genos_evee.csv` |
 | 07 | `07_hla_filter/` | 删除 GRCh38 HLA/MHC 区行（可 `--hla-filter no` 跳过） | **`vep_output.no_hla.csv`** |
 
 ---
@@ -338,7 +338,7 @@ pixi run bash scripts/run_full_pipeline.sh --help
 |------|------|
 | `<out-dir>/04_vep/vep_output.base.csv` | VEP 基础 CSV |
 | `<out-dir>/04_vep/raw_vep.tsv` | VEP 原始 TSV（默认保留） |
-| `<out-dir>/06_genos_evee_annotation/vep_output.with_genos_evee.csv` | GENOS-EVEE 宽表（HLA 过滤前） |
+| `<out-dir>/06_genos_evee_annotation/vep_output.with_genos_evee.csv` | GENOS-VarRisk 宽表（HLA 过滤前） |
 | `<out-dir>/07_hla_filter/vep_output.no_hla.csv` | **最终宽表**（默认） |
 | `<out-dir>/full_pipeline.outputs.tsv` | 各步产物路径索引 |
 | `<out-dir>/logs/full_pipeline.log` | 全流程日志 |
@@ -407,7 +407,7 @@ VEP 配置 [`modules/vep_runner/config/vep_runner_config.json`](modules/vep_runn
 |------|------|
 | [complete_pipeline/README.md](complete_pipeline/README.md) | API 与 CLI 详细参数 |
 | [modules/vep_runner/README.md](modules/vep_runner/README.md) | VEP 插件与注释库安装 |
-| [modules/genos_evee_annotation/README.md](modules/genos_evee_annotation/README.md) | GENOS-EVEE 注释与数据库构建 |
+| [modules/genos_evee_annotation/README.md](modules/genos_evee_annotation/README.md) | GENOS-VarRisk 注释与数据库构建 |
 | [modules/hla_filter/README.md](modules/hla_filter/README.md) | HLA/MHC 区行过滤 |
 | [modules/result_sorting/README.md](modules/result_sorting/README.md) | 可选致病性排序（主流程默认不启用） |
 | [resource_mock/README.md](resource_mock/README.md) | 联调用迷你 mock 库 |

modules/pipeline/complete_pipeline/README.md

@@ -3,7 +3,7 @@
 > 完整架构和数据资源下载说明见上一层 `../README.md`。本文件保留主程序/API 的详细调用说明。
 
 
-本目录包含 FastAPI 服务（`full_pipeline_api.py`）与 API 任务目录，负责把 liftover（可选）、phasing、VCF 前处理、假基因注释、VEP runner、VCF INFO 回填、GENOS-EVEE 注释、HLA 过滤串成一个完整流程。
+本目录包含 FastAPI 服务（`full_pipeline_api.py`）与 API 任务目录，负责把 liftover（可选）、phasing、VCF 前处理、假基因注释、VEP runner、VCF INFO 回填、GENOS-VarRisk 注释、HLA 过滤串成一个完整流程。
 
 主程序：
 
@@ -307,7 +307,7 @@ API 的字段和主程序参数一一对应。常规字段和高级覆盖字段
 | `top_k_transcripts` | `--top-k-transcripts` | 高级覆盖 | 每个 variant-gene 保留的转录本数量。 |
 | `clinical_tissue` | `--clinical-tissue` | 高级覆盖 | 手动传 GTEx tissue。 |
 | `keep_raw_vep` | `--keep-raw-vep` | 高级覆盖 | `true` 对应 `yes`，`false` 对应 `no`。 |
-| `genos_evee_db` | `--genos-evee-db` | 高级覆盖 | GENOS-EVEE CPRA 数据库路径。 |
+| `genos_evee_db` | `--GENOS-VarRisk-db` | 高级覆盖 | GENOS-VarRisk CPRA 数据库路径。 |
 | `dry_run` | `--dry-run` | 调试 | 只打印命令，不实际运行。 |
 
 示例：API 只跑 chr1：

modules/pipeline/complete_pipeline/full_pipeline_api.py

@@ -75,7 +75,7 @@ class RunRequest(BaseModel):
     java_bin: Optional[str] = Field(None, description="Advanced override: Java executable for Beagle")
     top_k_transcripts: Optional[int] = Field(None, description="Advanced override: transcript selection count")
     clinical_tissue: str = Field("", description="Advanced override: GTEx tissue name for phenotype-aware transcript expression")
-    genos_evee_db: Optional[str] = Field(None, description="Advanced override: indexed GENOS-EVEE CPRA TSV.GZ")
+    genos_evee_db: Optional[str] = Field(None, description="Advanced override: indexed GENOS-VarRisk CPRA TSV.GZ")
     keep_raw_vep: Optional[bool] = Field(None, description="Advanced override: keep raw VEP TSV")
     dry_run: bool = False
 
@@ -298,7 +298,7 @@ def add_option(name: str, value) -> None:
     add_option("--java-bin", resolve_path(req.java_bin) if req.java_bin else None)
     add_option("--top-k-transcripts", req.top_k_transcripts)
     add_option("--clinical-tissue", req.clinical_tissue)
-    add_option("--genos-evee-db", resolve_path(req.genos_evee_db) if req.genos_evee_db else None)
+    add_option("--GENOS-VarRisk-db", resolve_path(req.genos_evee_db) if req.genos_evee_db else None)
     if req.keep_raw_vep is not None:
         cmd.extend(["--keep-raw-vep", "yes" if req.keep_raw_vep else "no"])
     if req.dry_run:

modules/pipeline/modules/genos_evee_annotation/README.md

@@ -1,21 +1,21 @@
-# GENOS-EVEE 疾病预测数据库注释模块
+# GENOS-VarRisk 疾病预测数据库注释模块
 
-将疾病预测模型输出中的 `p_fusion` 按 CPRA 匹配到宽表，列名固定为 `GENOS-EVEE`。对应主流程第 **06** 步。
+将疾病预测模型输出中的 `p_fusion` 按 CPRA 匹配到宽表，列名固定为 `GENOS-VarRisk`。对应主流程第 **06** 步。
 
 ## 脚本
 
 | 脚本 | 用途 |
 |------|------|
 | `scripts/build_genos_evee_db.py` | 将 prediction TSV shard 构建为 bgzip + tabix 数据库 |
-| `scripts/add_genos_evee_to_csv.py` | 在 05 宽表中加入 `GENOS-EVEE` 列 |
+| `scripts/add_genos_evee_to_csv.py` | 在 05 宽表中加入 `GENOS-VarRisk` 列 |
 
 ## 主流程
 
 ```bash
 bash scripts/run_full_pipeline.sh \
   --input-vcf /path/to/input.vcf.gz \
   --out-dir /path/to/output \
-  --genos-evee-db /path/to/genos_evee.cpra.tsv.gz
+  --GENOS-VarRisk-db /path/to/genos_evee.cpra.tsv.gz
 ```
 
 默认数据库路径由 `.env` 的 `FULL_PIPELINE_GENOS_EVEE_DB` 或 `config/path_utils.py` 解析。数据库不存在时流程仍可运行，该列全部填 `-`。

modules/pipeline/modules/genos_evee_annotation/scripts/add_genos_evee_to_csv.py

@@ -9,7 +9,7 @@
 import tempfile
 from pathlib import Path
 
-OUTPUT_COLUMN = "GENOS-EVEE"
+OUTPUT_COLUMN = "GENOS-VarRisk"
 
 
 def normalize_chrom(value: str) -> str:
@@ -70,7 +70,7 @@ def annotate(input_csv: Path, database: Path | None, output_csv: Path, log_json:
     output_csv.parent.mkdir(parents=True, exist_ok=True)
     database_available = database is not None and database.is_file() and Path(f"{database}.tbi").is_file()
     if database is not None and database.exists() and not database_available:
-        raise FileNotFoundError(f"GENOS-EVEE database index not found: {database}.tbi")
+        raise FileNotFoundError(f"GENOS-VarRisk database index not found: {database}.tbi")
     if database_available and shutil.which("tabix") is None:
         raise RuntimeError("tabix command not found")
 
@@ -153,9 +153,9 @@ def annotate(input_csv: Path, database: Path | None, output_csv: Path, log_json:
 
 
 def main() -> None:
-    parser = argparse.ArgumentParser(description="Add GENOS-EVEE p_fusion scores to a V3 wide CSV before sorting.")
+    parser = argparse.ArgumentParser(description="Add GENOS-VarRisk p_fusion scores to a V3 wide CSV before sorting.")
     parser.add_argument("--input-csv", required=True)
-    parser.add_argument("--database", help="Indexed GENOS-EVEE .tsv.gz; missing/empty means fill '-'")
+    parser.add_argument("--database", help="Indexed GENOS-VarRisk .tsv.gz; missing/empty means fill '-'")
     parser.add_argument("--output-csv", required=True)
     parser.add_argument("--log-json")
     args = parser.parse_args()

modules/pipeline/modules/genos_evee_annotation/scripts/build_genos_evee_db.py

@@ -165,7 +165,7 @@ def build_database(inputs: list[Path], output: Path, threads: int, log_json: Pat
         )
         unique = 0
         with sorted_path.open("r", encoding="utf-8") as src, merged.open("w", encoding="utf-8") as dst:
-            dst.write("#CHROM\tPOS\tREF\tALT\tGENOS-EVEE\n")
+            dst.write("#CHROM\tPOS\tREF\tALT\tGENOS-VarRisk\n")
             previous: tuple[str, str, str, str] | None = None
             for line in src:
                 parts = line.rstrip("\n").split("\t")
@@ -190,7 +190,7 @@ def build_database(inputs: list[Path], output: Path, threads: int, log_json: Pat
 
 
 def main() -> None:
-    parser = argparse.ArgumentParser(description="Build indexed GENOS-EVEE CPRA database from prediction TSV shards.")
+    parser = argparse.ArgumentParser(description="Build indexed GENOS-VarRisk CPRA database from prediction TSV shards.")
     parser.add_argument("--input", action="append", required=True, help="TSV file, directory, or glob; repeatable")
     parser.add_argument("--output", required=True, help="Output .tsv.gz database")
     parser.add_argument("--threads", type=int, default=4)

modules/pipeline/modules/hla_filter/README.md

@@ -1,6 +1,6 @@
 # HLA/MHC 区域行过滤模块
 
-从 GENOS-EVEE 宽表中删除 GRCh38 HLA/MHC 区域的变异行。对应主流程第 **07** 步（默认开启）。
+从 GENOS-VarRisk 宽表中删除 GRCh38 HLA/MHC 区域的变异行。对应主流程第 **07** 步（默认开启）。
 
 ## 过滤区间
 

modules/pipeline/modules/result_sorting/README.md

@@ -4,7 +4,7 @@
 
 ## 与主流程的关系
 
-当前 `scripts/run_full_pipeline.sh` **默认不调用**本模块。主流程在 06 GENOS-EVEE 注释之后，由 07 HLA 过滤产出最终宽表：
+当前 `scripts/run_full_pipeline.sh` **默认不调用**本模块。主流程在 06 GENOS-VarRisk 注释之后，由 07 HLA 过滤产出最终宽表：
 
 ```text
 05 vcf_info_to_csv → 06 genos_evee_annotation → 07 hla_filter（默认最终 CSV）
@@ -14,7 +14,7 @@
 
 ## 输入
 
-- `--input-csv`：带 `vcf_info_*`（及可选 `GENOS-EVEE`）列的宽表，通常来自 `05_vcf_info_to_csv/` 或 `06_genos_evee_annotation/`。
+- `--input-csv`：带 `vcf_info_*`（及可选 `GENOS-VarRisk`）列的宽表，通常来自 `05_vcf_info_to_csv/` 或 `06_genos_evee_annotation/`。
 - `--output-csv`：排序后的 CSV。
 
 ## 单独运行示例

modules/pipeline/modules/vcf_info_to_csv/README.md

@@ -2,7 +2,7 @@
 
 功能：读取进入 04 VEP runner 的 VCF，把 VCF 中所有 INFO 字段展开为 `vcf_info_<INFO_ID>` 列，并把样本 FORMAT 展开为 `vcf_format_*` 列，追加到 04 生成的 VEP CSV 中。
 
-主流程中本模块为第 **05** 步；输出 `vep_output.with_info.csv` 进入 **06** GENOS-EVEE 注释，再经 **07** HLA 过滤（默认）得到最终宽表。
+主流程中本模块为第 **05** 步；输出 `vep_output.with_info.csv` 进入 **06** GENOS-VarRisk 注释，再经 **07** HLA 过滤（默认）得到最终宽表。
 
 ## 输入
 

modules/pipeline/modules/vep_runner/scripts/run_vep_to_csv.py

@@ -805,6 +805,50 @@ def parse_uploaded_variation(value: str) -> dict[str, str]:
     return parsed
 
 
+def unique_ordered(values: list[str]) -> list[str]:
+    seen: set[str] = set()
+    result: list[str] = []
+    for value in values:
+        if not value or value in seen:
+            continue
+        seen.add(value)
+        result.append(value)
+    return result
+
+
+def split_variant_ids(variant_id: str) -> list[str]:
+    variant_id = (variant_id or "").strip()
+    if not variant_id or variant_id in {".", "-"}:
+        return []
+    return unique_ordered([variant_id, *[item.strip() for item in variant_id.split(";")]])
+
+
+def parse_vep_location(location: str) -> tuple[str, str] | None:
+    location = (location or "").strip()
+    if not location or ":" not in location:
+        return None
+    chrom, _, interval = location.partition(":")
+    start = interval.split("-", 1)[0].strip()
+    if not chrom or not start:
+        return None
+    return chrom, start
+
+
+def location_lookup_keys(chrom: str, pos: str, ref: str, alt: str) -> list[str]:
+    keys = [f"loc:{chrom}:{pos}:{alt}"]
+    try:
+        shifted_pos = str(int(pos) + 1)
+    except ValueError:
+        return keys
+
+    keys.append(f"loc:{chrom}:{shifted_pos}:{alt}")
+    if len(ref) < len(alt) and alt.startswith(ref):
+        keys.append(f"loc:{chrom}:{shifted_pos}:{alt[len(ref):] or '-'}")
+    elif len(ref) > len(alt) and ref.startswith(alt):
+        keys.append(f"loc:{chrom}:{shifted_pos}:-")
+    return keys
+
+
 def vep_uploaded_variation_keys(chrom: str, pos: str, ref: str, alt: str) -> list[str]:
     keys = [f"{chrom}_{pos}_{ref}/{alt}"]
     if len(ref) == len(alt):
@@ -838,9 +882,52 @@ def original_variant_lookup_keys(
     back to the original VCF allele.
     """
     keys = vep_uploaded_variation_keys(chrom, pos, ref, alt)
-    if variant_id and variant_id not in {".", "-"}:
-        keys.append(variant_id)
-    return keys
+    keys.extend(location_lookup_keys(chrom, pos, ref, alt))
+    for item in split_variant_ids(variant_id):
+        keys.append(f"id:{item}")
+        keys.append(f"id:{item}|allele:{alt}")
+    return unique_ordered(keys)
+
+
+def vep_row_candidate_keys(row: dict[str, str]) -> list[str]:
+    uploaded = row.get("Uploaded_variation", "")
+    allele = row.get("Allele", "")
+    keys: list[str] = []
+    if uploaded:
+        keys.append(uploaded)
+
+    parsed = parse_uploaded_variation(uploaded)
+    chrom = parsed.get("chrom", "")
+    pos = parsed.get("pos", "")
+    ref = parsed.get("ref", "")
+    alt = parsed.get("alt", "")
+    if chrom and pos and ref and alt:
+        keys.extend(vep_uploaded_variation_keys(chrom, pos, ref, alt))
+        if allele and allele != alt:
+            keys.extend(vep_uploaded_variation_keys(chrom, pos, ref, allele))
+
+    location = parse_vep_location(row.get("Location", ""))
+    if location and allele:
+        loc_chrom, loc_pos = location
+        keys.append(f"loc:{loc_chrom}:{loc_pos}:{allele}")
+
+    for item in split_variant_ids(uploaded):
+        if allele:
+            keys.append(f"id:{item}|allele:{allele}")
+        if alt:
+            keys.append(f"id:{item}|allele:{alt}")
+        keys.append(f"id:{item}")
+        keys.append(item)
+
+    return unique_ordered(keys)
+
+
+def lookup_original_variant(row: dict[str, str], original_variants) -> dict[str, str] | None:
+    for key in vep_row_candidate_keys(row):
+        value = original_variants.get(key)
+        if value is not None:
+            return value
+    return None
 
 
 def vcf_info_column_name(info_id: str) -> str:
@@ -1915,7 +2002,7 @@ def convert_vep_table_to_csv(
             row = dict(zip(header, parts))
             extra = parse_extra(row.pop("Extra", ""))
             row.update(extra)
-            original_variant = original_variants.get(row.get("Uploaded_variation", ""))
+            original_variant = lookup_original_variant(row, original_variants)
             add_requested_columns(row, extra, original_variant)
             extra_keys.update(extra)
             variant_id = row.get("Uploaded_variation", "")
@@ -2026,7 +2113,7 @@ def parse_vep_table_groups(
             extra = parse_extra(row.pop("Extra", ""))
             row.update(extra)
             variant_id = row.get("Uploaded_variation", "")
-            original_variant = original_variants.get(variant_id)
+            original_variant = lookup_original_variant(row, original_variants)
             add_requested_columns(row, extra, original_variant)
             if current_key is not None and variant_id != current_key:
                 yield current_key, current_rows

modules/pipeline/resource_mock/README.md

@@ -25,7 +25,7 @@ bash scripts/run_full_pipeline.sh \
   --phasing no \
   --ccre-bed resource_mock/regulatory/hg38/encode_screen_v4_grch38_ccre.slim.mock.bed.gz \
   --ncrna-bed resource_mock/ncrna/hg38/gencode.v49.ncrna_gene.slim.mock.bed.gz \
-  --genos-evee-db resource_mock/genos_evee/genos_evee.cpra.mock.tsv.gz \
+  --GENOS-VarRisk-db resource_mock/genos_evee/genos_evee.cpra.mock.tsv.gz \
   --dry-run
 ```