BIOP02-148: Table 1 코호트 특성표 5암종 + participant flow - #138
Merged
Conversation
5암종(유방·폐·대장·위·두경부) 환자수/슬라이드수/train-holdout/나이/성별/ stage/grade/TSS site수 + 엔드포인트별 유병률/결측/assay source. 전부 기존 정본 파일(patient_labels.csv/split.csv/manifest/mil_cost_results.json) 집계, 신규 라벨 없음. grade 커버리지 유방·폐·대장 0%(cBioPortal 속성 자체 부재, 실측 확인) vs 위·두경부 ~100% 그대로 노출. check_table1_numbers.py로 자체 검증: holdout n_pos 전부 mil_cost_results.json 과 1:1 일치 + manuscript/sections/02_results.md R1/R2 인용 수치와도 불일치 0건(DoD "분모 불일치 0건" 항목). Table 1 배치는 본문 권장이나 원고 prose는 미수정(리뷰 후 편입 제안). 독립 리뷰 요청 대상: 박세진(sezinie000) 또는 지용기, DoD 명시.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
BIOP02-148 — Table 1(코호트 특성표) 5암종
claim_level: descriptive(가설 검정 아님, 사전등록 대상 없음)범위
티켓 본문 전체 범위(축소 옵션 미사용) — 5암종(유방·폐·대장·위·두경부) × 환자수/슬라이드수/train·holdout/나이/성별/stage/grade/tissue source site 수 +
엔드포인트별 유병률·결측·assay source + participant flow. 전부 기존 정본 파일(
patient_labels.csv/split.csv/tcga_brca_manifest.csv/mil_cost_results.json)집계 — 신규 라벨·모델링 없음.
산출물
fetch_table1_clinical.py→table1_clinical.json— age/sex/stage/grade/TSS site (cBioPortal). 정정 1건: GRADE·TISSUE_SOURCE_SITE_CODE가 PATIENT-level이아니라 SAMPLE-level 속성임을 최초 실행에서 0건으로 확인 후 발견 → 재실행.
build_table1.py→table1_master.json(정본) +TABLE1.md+PARTICIPANT_FLOW.mdcheck_table1_numbers.py— DoD "결과 파일과 1:1 대조" 자체 검증핵심 발견
GRADE속성 자체가 없음 — 체리피킹아니라 실측 확인, BIOP02-140에서 폐도 동일 확인). 숨기지 않고 표에 그대로 노출.
LUNG_NSCLC/full/SUBTYPE_BASELINE_NOTE.md(기존 산출물)가 R1이 인용한 "조직형만으로 0.793" 기준선의 정확한 출처임을 확인. 리뷰 대기 중인BIOP02-140(BIOP02-140 v1: 폐 공변량 기준선 + KRAS/EGFR 증분가치 — kras_g12c 🔴 강등 권고 #136)이 이 기준선의 확장판(0.802)이라 같은 결론을 독립적으로 강화(참고용, 본 PR 범위 밖).
DoD "본문·표 R1/R2와 분모 불일치 0건" 검증
check_number_drift.py(BIOP02-107)는 AUROC류·3개 문서만 봐서 Table 1은 범위 밖 — 별도로check_table1_numbers.py작성해manuscript/sections/02_results.md의R2 표 6행 + R1 prose 헤드라인 인용 3건을
table1_master.json과 대조. 불일치 0건.Table 1 배치
본문(main text) 권장 — TRIPOD+AI 13b 표준,
02_results.mdR1 앞이 자연스러움. 다만 원고 prose는 이번 PR에서 직접 수정하지 않음 — 다른 분들이 쓰고 계신원고라 배치·문구는 리뷰 후 합의해서 별도 커밋으로 편입 제안.
DoD
관련: Jira BIOP02-148, BIOP02-76(TRIPOD 매핑) · BIOP02-138(#3 operating-point) · BIOP02-140(같은 KRAS
기준선 교차확인)