This lesson is only available in Korean for now.
Dr. Son's Perturb-seq, First Steps

Silence one gene —
what does the immune cell tell you?

CD4+ T cells from 4 healthy donors, each gene silenced one at a time by CRISPRi, the remaining 20,000 read out by single-cell RNA-seq. The result: 11,526 genes × 3 conditions × 2.1 million cells.
Over these 20 lessons you handle the data yourself — differential expression, functional clusters, GWAS disease mapping, and drug target selection. Python runs right inside this page.

🧬 CRISPRi🔬 single-cell🩸 CD4+ T cells 📊 DE analysis🕸️ gene clusters💊 drug targets🐍 Python
Start lesson 1 →

How you'll learn

🔬With the real data

Not a stand-in. The 33,983-row DE statistics, 112 cluster annotations, and 17 autoimmune disease enrichment results from Zhu et al. (2026), exactly as published.

🧪Design decisions first

Why primary cells not cell lines, why CRISPRi not Cas9, why three conditions — the rationale behind each design choice, before touching the data.

⚠️Pitfalls first

Changing the condition flips the answer, too few cells shake the conclusion, you silence the neighbour gene too. You feel how you get fooled before trusting results.

🛠️Learning by handling it

Every lesson has an interactive widget — a knockdown effect slider, a cluster explorer, a disease enrichment heatmap, a drug target comparison, all turned by you.

Full curriculum

01
왜 Perturb-seq인가 — 유전자 하나를 끄면 2만 개가 대답한다
Zhu, Dann et al. (2026) Fig. 1a — 스크린 개요
✓
02
크리스퍼의 두 얼굴 — 자르는 것과 끄는 것
Zhu, Dann et al. (2026) Methods — CRISPRi (dCas9-KRAB)
✓
03
CD4+ T세포 — 면역계의 지휘관
Zhu, Dann et al. (2026) Introduction — primary human T cells
✓
04
세 가지 조건 — 휴식, 8시간, 48시간
Zhu, Dann et al. (2026) Fig. 1b — condition-specific effects
✓
05
단일세포 RNA-seq — 세포 하나의 레시피를 읽는다
Zhu, Dann et al. (2026) Methods — single-cell profiling
✓
06
스크린 설계 — 11,526개 유전자를 한 번에
Zhu, Dann et al. (2026) Fig. 1a,b — library design
✓
07
차등 발현 분석 — 무엇이 바뀌었나를 센다
Zhu, Dann et al. (2026) SI Table S1 — DE statistics
✓
08
유사벌크 — 왜 세포를 합쳐야 하나
Zhu, Dann et al. (2026) Methods — pseudobulk DE with DESeq2
✓
09
표적 효과 — 진짜로 꺼졌나 확인한다
Zhu, Dann et al. (2026) SI Table S1 — ontarget_effect_size
✓
10
가이드 일치 — 두 가이드가 같은 말을 하는가
Zhu, Dann et al. (2026) SI — guide concordance analysis
✓
11
【실습】 가장 강한 교란 — TADA2B에서 CTLA4까지
Zhu, Dann et al. (2026) SI Table S1 — top hits by n_total_de_genes
✓
12
유전자 팀 — 비슷하게 끊어지면 같은 팀이다
Zhu, Dann et al. (2026) Fig. 3 — functional gene clusters
✓
13
【실습】 112개 클러스터를 탐험한다
Zhu, Dann et al. (2026) SI Table S4 — cluster annotations
✓
14
함정 — 조건을 바꾸면 답이 바뀐다
Zhu, Dann et al. (2026) Fig. 2 — condition-specific perturbation effects
✓
15
함정 — 세포가 적으면 결론이 흔들린다
Zhu, Dann et al. (2026) Methods — cell number thresholds
✓
16
함정 — 오프타깃, 옆 유전자를 같이 끄다
Zhu, Dann et al. (2026) Methods — off-target filtering
✓
17
질병 연결 — GWAS 주소에 보안 카메라를 단다
Zhu, Dann et al. (2026) Fig. 4 — autoimmune disease enrichment
✓
18
약물 표적 — 좋은 표적과 나쁜 표적
Zhu, Dann et al. (2026) Discussion — translational implications
✓
19
【실습 종합】 논문 그림을 처음부터 그린다
Zhu, Dann et al. (2026) Fig. 3,4 — cluster heatmap & disease enrichment
✓
20
정리 — 이 데이터로 말할 수 있는 것과 없는 것
Zhu, Dann et al. (2026) Discussion
✓