Ep180: John Rinn on Discovering Drugs Based on Long Non-Coding RNA

Episode Summary

Executive Summary: John Wren traces his path from a nontraditional Minnesota upbringing to becoming a leading researcher in long non-coding RNA (lncRNA), arguing that much of the “dark genome” is functional and medically important. He describes how new genomic tools, sequencing, computation, and AI revealed hidden RNA biology, enabled mechanistic studies and mouse knockouts, and now support Link Switch Therapeutics’ effort to develop RNA replacement therapies for disease-linked lncRNAs such as Renew syndrome.

Main Topics: From athlete to scientist (Priority: 4/5): Wren describes a childhood centered on hockey, running, snowboarding, and a non-linear academic path that eventually led him from community college to chemistry at the University of Minnesota Duluth. Discovering genomics and the dark genome (Priority: 5/5): At Yale, Wren shifted from crystallography to genomics after seeing unresolved debates about gene number and realizing data could answer questions about the genome’s unexplored regions. Long non-coding RNAs as functional biology (Priority: 5/5): The conversation explains how lncRNAs were initially dismissed as noise or junk DNA, but evidence from Hox clusters, epigenetic regulation, and X-chromosome biology suggested real function. Technology as the driver of discovery (Priority: 5/5): Microarrays, sequencing, epigenetic profiling, computation, and now AI repeatedly expanded what could be measured, validated, and interpreted in RNA biology. Proof of function through genetics and mouse models (Priority: 5/5): Wren recounts moving from expression maps to knockout mice and disease genetics, using phenotypes and inherited mutations to argue that some lncRNAs are truly causal. Therapeutic translation via Link Switch Therapeutics (Priority: 5/5): He explains the startup’s goal: use RNA itself as a drug to restore or switch diseased lncRNA function, prioritizing genetically validated targets like Renew syndrome. Colorado, collaboration, and teaching (Priority: 3/5): Wren says Boulder offered a unique RNA ecosystem, shared tools, and rigorous colleagues; he also emphasizes his free online bioinformatics teaching as a way to democratize discovery biology.

Key Arguments: The genome’s non-coding majority is not merely junk; a meaningful subset contains functional lncRNAs that regulate biology and disease. Scientific skepticism was useful because it forced stronger validation: orthogonal assays, sequencing, epigenetic signatures, and mouse knockouts. Microarrays and later sequencing were essential because they allowed researchers to look beyond known genes and tile whole genomic regions. lncRNAs can act through epigenetic mechanisms and in some cases behave like trans-acting functional molecules rather than passive transcripts. Human genetics increasingly supports lncRNA causality, with disease-linked mutations showing that some lncRNAs are bona fide therapeutic targets. RNA replacement may be a viable therapeutic modality, analogous to protein replacement therapies, especially as delivery technologies improve. AI will be most useful if it is trained on carefully designed, high-signal RNA biology data rather than generic scraped text. Discovery biology remains highly relevant because the field is still learning what is functional in the dark genome and how to prioritize targets.

Data Points: Protein-coding fraction of genome: ~2% - Wren notes that roughly 98% of the genome does not encode proteins. Non-coding fraction of genome: ~98% - Described as the unexplored “dark genome” or long non-coding space. Gene-number debate in early genomics: 20,000 to 100,000 genes - He recalls the lack of data and wide speculation during the human genome era. First microarray scale: ~6,000 dots - Early yeast-era arrays used dot spots representing genomic elements. Whole-chromosome microarray used in lab: 20,000 spots - His Yale work tiled chromosome 22 rather than only known genes. Long non-coding RNA length definition: >200 nucleotides - He explains this is largely an extraction-column artifact rather than a true biological cutoff. Hox cluster size reference: ~100 kilobases - Used to illustrate how a small genomic region can control major developmental programs. Number of lncRNAs identified in an early paper: 174 - He cites an early Hox-related discovery paper from his postdoc era. Gene-like filtered lncRNAs at Broad: ~3,000 - After integrating promoter and histone-mark evidence, they narrowed candidates from broader noise. Knockout mice made: 20 - A high-throughput in vivo test of whether candidate lncRNAs had phenotypes. Disease-linked lncRNA loci mentioned: 4 - He says there are now four Mendelian lncRNA disease loci, naming Fire, Chaser, Fender, and U4. Renew syndrome burden: 100,000 predicted children - He cites a newly discussed lncRNA-related neurodevelopmental disorder. Course length: <20 hours - His free end-to-end bioinformatics course can be completed in less than 20 hours. Course reach: ~1,000 people - He says the online course has attracted about a thousand learners on YouTube.

Pivotal Quotes: "“If it has to be a gene, look like a gene, smell like a gene, taste like a gene, bark like a gene.”" — John Wren: Describing the 2009-style filtering logic used to validate candidate long non-coding RNAs. "“I think the RNA is the drug.”" — John Wren: Explaining Link Switch Therapeutics’ core therapeutic philosophy. "“If we build it, it will come.”" — John Wren: Summarizing his belief that the right platform and data will enable future RNA medicines.

Implications: The episode argues that lncRNA biology is moving from curiosity to medicine: better tools, genetics, and AI are exposing new targets, and RNA replacement could open a new therapeutic class if delivery and validation continue to improve.