Lex Fridman PodcastManolis Kellis: Biology of Disease | Lex Fridman Podcast #133
Lex Fridman and Manolis Kellis on decoding Human Disease: Genetics, Brain Circuits, and Future Therapies.
In this episode of Lex Fridman Podcast, featuring Lex Fridman and Manolis Kellis, Manolis Kellis: Biology of Disease | Lex Fridman Podcast #133 explores decoding Human Disease: Genetics, Brain Circuits, and Future Therapies Lex Fridman and Manolis Kellis explore how modern human genetics and computational biology are transforming our understanding of complex diseases such as obesity, Alzheimer’s, schizophrenia, heart disease, and metabolic disorders.
At a glance
WHAT IT’S REALLY ABOUT
Decoding Human Disease: Genetics, Brain Circuits, and Future Therapies
- Lex Fridman and Manolis Kellis explore how modern human genetics and computational biology are transforming our understanding of complex diseases such as obesity, Alzheimer’s, schizophrenia, heart disease, and metabolic disorders.
- Kellis explains the shift from traditional one-gene, animal-model biology to large-scale human genomics, where millions of natural genetic perturbations across thousands of people and phenotypes reveal causal mechanisms of disease.
- They dive into multi-layered biological circuitry—from DNA variants, epigenomics, and gene expression to cell types, organs, and behavior—and show how convergent pathways (like calcium signaling, immune function, and energy metabolism) emerge across many diseases.
- The discussion highlights powerful new tools (CRISPR, single-cell sequencing, high-throughput assays, AI-driven analysis) that enable systematic mapping of disease circuitry and point toward a coming era of precision, systems-level, and multi-target therapeutics.
IDEAS WORTH REMEMBERING
7 ideasHuman genetics has flipped the old model of biology on its head.
Instead of learning basic mechanisms in mice and then mapping them to humans, we now use the vast diversity of human genetic variants and natural 'experiments' to discover causal genes, pathways, and tissues, which then drive basic biological insight.
Most disease variants act through gene regulation, not by breaking proteins.
About 93% of disease-associated variants lie outside protein-coding regions, mainly in regulatory elements (enhancers), so understanding long-range genome circuitry—what variants control which genes in which cell types—is essential for pinpointing mechanisms and drug targets.
Complex diseases are polygenic but converge on a limited set of pathways.
Thousands of small-effect variants and regulatory elements may differ between people, but they often funnel into common processes (e.g., calcium signaling in schizophrenia, immune microglia in Alzheimer’s, lipid metabolism and thermogenesis in obesity), making pathway-level interventions feasible.
Multi-level data integration is key to decoding disease mechanisms.
By linking genetic variants to epigenomic marks, gene expression, single-cell profiles, cell-to-cell communication, organ-level measures, and clinical phenotypes, researchers can trace full causal chains from a nucleotide change to molecular, cellular, and behavioral outcomes.
New experimental platforms let scientists test thousands of hypotheses in parallel.
Technologies such as massively parallel reporter assays (MPRA), high-throughput CRISPR perturbations, and single-cell RNA/ATAC sequencing enable simultaneous testing of tens of thousands of variants, enhancers, and genes, dramatically accelerating the mapping of disease circuitry.
The FTO obesity locus illustrates how deep circuitry analysis enables interventions.
What was thought to be an obesity gene (FTO) turned out to be a regulatory region controlling distant genes IRX3/IRX5 in fat progenitors, tipping cells between energy-burning (thermogenic) and energy-storing fat; a single base change can flip cellular behavior, revealing multiple potential therapeutic levers.
The future of medicine is systems-level, personalized, and multi-target.
Kellis envisions treatments that consider an individual’s common, rare, and somatic variants, their molecular and clinical profiles, and then use combinatorial interventions (DNA/RNA drugs, cell-type-specific constructs, small molecules) to modulate whole networks rather than single genes, while minimizing on-target side effects.
WORDS WORTH SAVING
5 quotesUnderstanding human disease is the most complex challenge in modern science.
— Manolis Kellis
You cannot solve disease with traditional biology. You have to think genomically.
— Manolis Kellis
This is a paper about one nucleotide in the human genome… one bit of information.
— Manolis Kellis
The confluence of technologies, computation, data, insight, and tools for manipulation is unprecedented in human history.
— Manolis Kellis
Disease is gonna be fundamentally altered and alleviated as we go forward.
— Manolis Kellis
QUESTIONS ANSWERED IN THIS EPISODE
5 questionsHow should we ethically handle personal genetic risk information, given that it can shape life choices but also cause anxiety or discrimination?
Lex Fridman and Manolis Kellis explore how modern human genetics and computational biology are transforming our understanding of complex diseases such as obesity, Alzheimer’s, schizophrenia, heart disease, and metabolic disorders.
If most disease variants act through regulation rather than coding changes, how might that change drug discovery and target prioritization strategies in pharma?
Kellis explains the shift from traditional one-gene, animal-model biology to large-scale human genomics, where millions of natural genetic perturbations across thousands of people and phenotypes reveal causal mechanisms of disease.
What are the biggest computational bottlenecks in integrating multi-omic and single-cell data into actionable clinical insights?
They dive into multi-layered biological circuitry—from DNA variants, epigenomics, and gene expression to cell types, organs, and behavior—and show how convergent pathways (like calcium signaling, immune function, and energy metabolism) emerge across many diseases.
How could systems-level, multi-gene interventions be tested safely in humans, given the potential for complex and unforeseen network effects?
The discussion highlights powerful new tools (CRISPR, single-cell sequencing, high-throughput assays, AI-driven analysis) that enable systematic mapping of disease circuitry and point toward a coming era of precision, systems-level, and multi-target therapeutics.
To what extent might future therapies blur the line between treating disease and enhancing normal human traits like cognition, stamina, or metabolism?
EVERY SPOKEN WORD
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome