←  Back to home
Research · in depth

How the genome's shape writes the rules of health and disease.

Our lab asks a deceptively simple question: why does the same genome behave so differently from one cell to the next? The answer lies partly in three dimensions — in how DNA is folded. Below is the longer story, and the four directions we are pursuing.

First, a 30-second primer

A single human cell holds roughly two meters of DNA, packed into a nucleus so small that thousands would fit across a pinhead. To fit, the genome folds — but not randomly.

Folding is guided by structural proteins (CTCF, cohesin) and a web of epigenetic marks. It pulls distant stretches of DNA into loops and neighborhoods called TADs, deciding which genes can reach their regulatory switches.

Change the fold, and you can change which genes turn on — and therefore what a cell becomes. When folding goes wrong, disease can follow.

Kim LabKim Lab, KAIST — jihunkim.kaist.ac.kr FOLD → CONTACT MAP Loops bring distant DNA together… …and show up as bright spots on the Hi-C map.
01
Cancer · genome instability
Diseases in focussolid tumors with structural variantsleukemiacopy-number–driven cancers

Unstable genomes in cancer, read in 3D

The problem
Infection, nuclear instability, and other environmental insults trigger somatic changes in the genome — point mutations, copy-number variation, structural variants — that are among the leading causes of cancer.
The gap
These variations are directly connected to changes in 3D genome networks and epigenetic landscapes — yet the molecular mechanisms that drive them remain poorly understood.
Our approach
We build 3D genome folding maps of cancer samples with unstable genome structures, then dissect the key mechanisms behind them and what they mean for the clinic.
cancerSNP · CNV · SV3D contact mapsclinical implication
Kim LabKim Lab, KAIST — jihunkim.kaist.ac.kr NORMAL UNSTABLE
Instability rewires the contact map — new bright blocks mark rearranged regions
02
Cancer · genome stability
Diseases in focusDNA-repair–deficient cancersreplication-stress tumorsgenome-instability syndromes

DNA damage, repair, and the broken fold

The problem
Our DNA is broken and repaired constantly. When repair fails, the errors that remain are a root cause of cancer — and repair does not happen on bare DNA, but on DNA that is folded into a complex 3D structure.
The gap
How the genome reorganizes its fold around a break, and whether that architecture helps or hinders faithful repair, is still poorly understood.
Our approach
We map how 3D folding changes at sites of DNA damage and during repair, and ask what goes wrong — structurally — when a cell takes the path toward cancer.
cancerDNA damage & repairgenome stability
Kim LabKim Lab, KAIST — jihunkim.kaist.ac.kr INTACT BREAK FOLD REORGANIZES & REPAIR PROTEINS GATHER
Around a break, chromatin refolds and repair factors converge — we ask how that shape shapes the outcome
03
Rare disease · repeat expansion
Diseases in focusspinocerebellar ataxiamyotonic dystrophyFriedreich's ataxia

When a repeat expansion refolds the genome

The problem
In a family of rare disorders — the spinocerebellar ataxias, myotonic dystrophy, Friedreich's ataxia and related conditions — a short stretch of DNA is repeated far too many times. Unlike the variants in our cancer work, these are inherited from birth.
The gap
An expanded repeat can nucleate repressive chromatin that spreads well beyond itself, quieting genes that carry no mutation at all. What decides how far that silencing travels, and how much of the disease follows from it, is still open.
Our approach
We map 3D folding in patient-derived and engineered models, ask which structural changes actually cause the expression defects, and test whether restoring the fold restores function.
rare diseaseneurologicalrepeat expansionpatient-derived models
Kim LabKim Lab, KAIST — jihunkim.kaist.ac.kr SHORT REPEAT repeat gene ON expansion EXPANDED REPEAT gene OFF
The repeat is small; the repressive chromatin it nucleates is not — and it reaches genes that carry no mutation at all
04
Technology · genome engineering
Where it appliesoncogene enhancer hijackingrepeat-expansion locisilenced disease genes

Learning to engineer 3D genome folding

The problem
The 3D genome is organized hierarchically — loops, TADs, compartments, chromosomes. The loop-extrusion model explains how loops form, but the cause-and-effect between loops, transcription, and epigenetic features is still unresolved.
The gap
Even how the genome's domain structures form in the first place remains elusive — which makes it hard to prove what folding actually does.
Our approach
We are developing new technology that can reshape 3D genome folding domains on demand, then test their function at target genes — turning correlation into causation.
genome engineeringloop extrusionsynthetic biologycause & effect
Kim LabKim Lab, KAIST — jihunkim.kaist.ac.kr BEFORE enhancer gene off engineer a loop AFTER gene ON ✓
We build a loop where there wasn't one — and watch the target gene switch on

Why it matters

3D genome folding is woven into the mechanisms of many diseases — cancer and leukemia, medulloblastoma, schizophrenia, repeat-expansion disorders, immune dysregulation, and metabolic change — as well as fundamental biology like development. By understanding the rules of folding, and learning to rewrite them, we aim to turn a basic mechanism into a lever for health.

That is why genome engineering sits alongside the disease work rather than after it. A fold you can rewrite on demand is a fold whose function you can prove — and a mechanism you can prove is one you can eventually treat.

Curious enough to fold in?

We take students and researchers from biology, medicine, bioengineering, and computer science. No prior 3D-genome experience required — just curiosity and drive.

See open positions →