Genetics Research Breakthroughs

💬 "If you don't have celiac disease, you don't need to avoid gluten." If only you had a dime for every time a healthcare provider said this. It turns out that celiac disease isn't the only autoimmune condition that can be caused or exacerbated by gluten consumption. 🧬 Celiac disease is considered a hereditary genetic condition, characterized by the presence of certain genetic alleles such as HLA-DQ2 and HLA-DQ8. But several other autoimmune diseases are also linked to gluten sensitivity, including type 1 diabetes, rheumatoid arthritis, and Sjögren's syndrome. 🍏 Nutrigenomics provides a genetic understanding for how common dietary components, such as gluten, affect our health and disease status. Let’s examine the interplay between gluten and various genes associated with autoimmune conditions. The genes implicated in these conditions fall into three main categories: 1️⃣ HLA genes (in pink): These are part of our immune system and play a crucial role in how our bodies recognize and respond to foreign substances, including gluten. Examples include: 📌 HLA-DQ2 and HLA-DQ8: Strongly associated with celiac disease 📌 HLA-DR3 and HLA-DR5: Linked to autoimmune thyroid diseases 2️⃣ Non-HLA genes (in blue): These include genes involved in immune regulation, intestinal barrier function, and cellular processes that can influence autoimmune responses. Examples include: 📌 IL-2 and IL-21: Involved in regulating immune responses 📌 INS: Insulin gene, associated with type 1 diabetes 📌 FOXP3: Important for the function of regulatory T cells 3️⃣ Shared genes (in orange): These genes are associated with multiple autoimmune conditions, suggesting common pathways in autoimmune dysfunction. Examples include: 📌 CTLA4: Regulates T cell responses & linked to several autoimmune diseases 📌 STAT4: Involved in immune cell signaling & associated with multiple autoimmune conditions 📌 MYO9B: Affects intestinal permeability & linked to celiac disease, rheumatoid arthritis, and lupus While not everyone with these genes will develop gluten sensitivity or an autoimmune condition, this research is a reminder that nutrition isn't one-size-fits-all, and some individuals might benefit from reducing gluten intake even in the absence of celiac diagnosis. Always consult with a healthcare professional before making significant dietary changes, but don't be afraid to advocate for yourself if you suspect gluten sensitivity. Your body's response to food is unique, and recognizing that is key to improving health outcomes.

Swedish researchers discovered enzyme repairing damaged DNA preventing cancer mutations forming naturally. Scientists at Karolinska Institute identified a DNA repair enzyme called APEX2 that can fix double-strand breaks—the most dangerous DNA damage type—with perfect accuracy, preventing the mutations that initiate cancer. Enhancing this enzyme's activity could prevent most cancers before they start. Every cell in your body suffers thousands of DNA damage events daily from UV radiation, metabolic byproducts, and environmental toxins. Most damage is repaired perfectly by cellular machinery, but double-strand breaks—where both DNA strands snap—are difficult to fix. Error-prone repair creates mutations; if these affect genes controlling cell division, cancer begins. It's estimated that DNA repair failure causes 90% of cancers. APEX2 is a backup repair enzyme that activates when primary repair systems fail. It precisely realigns broken DNA ends and recruits ligase enzymes to reconnect strands without errors. Swedish scientists found that people with naturally higher APEX2 activity have significantly lower lifetime cancer risk. They developed a gene therapy that increases APEX2 production in tissues most vulnerable to cancer—colon, breast, lung, prostate. Animal studies showed dramatic results: mice engineered to develop cancer rapidly remained cancer-free when APEX2 was enhanced. Even when exposed to carcinogens, their DNA repair prevented mutation accumulation. Human trials are beginning with people carrying genetic mutations that predispose to cancer. If successful, we're potentially discovering cancer prevention at the most fundamental level—stopping the initial DNA damage that begins the disease process. Source: Karolinska Institute, Nature Genetics 2025 #DNARepair #CancerPrevention #GeneticMedicine #MutationPrevention #CellularBiology #OncologyResearch

Synthetic biology is - quite literally - our future. A goundbreaking new biological foundation model Evo2 achieves state-of-the-art prediction of genetic variation impacts and generates coherent genome sequences, spanning all domains of life. A diverse team from leading research institutions including Arc Institute Stanford University NVIDIA University of California, Berkeley trained the model on 9.3 trillion DNA base pairs and has fully shared all code, parameters, and data. A few highlights from the paper (link in comments) 🔬 Zero-shot prediction achieves state-of-the-art accuracy in genetic variant interpretation. Evo 2 can predict the functional consequences of genetic mutations across all domains of life without specialized training. It surpasses existing models in assessing the pathogenicity of both coding and noncoding variants, including BRCA1 cancer-linked mutations. This generalist capability suggests Evo 2 could revolutionize genetic disease research, reducing reliance on expensive, manually curated datasets. 🛠 Genome-scale generation paves the way for synthetic life design. Evo 2 can generate full-length genome sequences with realistic structure and function, including mitochondrial genomes, bacterial chromosomes, and yeast DNA. Unlike prior models, Evo 2 ensures natural sequence coherence, improving synthetic biology applications like engineered microbes or artificial organelles. This sets the stage for programmable biology at an unprecedented scale. 🧬 Unprecedented long-context understanding revolutionizes genomic analysis. Evo 2 operates with a context window of up to 1 million nucleotides—far beyond the capabilities of previous models—allowing it to analyze genomic features across vast distances. This ability enables it to accurately identify regulatory elements, exon-intron boundaries, and structural components critical for understanding genome function. Its long-context recall is a major breakthrough for interpreting complex biological sequences. 🎛 Inference-time search enables controllable epigenomic design. Evo 2’s generative abilities extend beyond raw DNA sequence to epigenomic features, allowing researchers to design sequences with specific chromatin accessibility patterns. This approach successfully encoded Morse code messages into synthetic epigenomes, demonstrating a new method for controlling gene regulation via AI. This could lead to breakthroughs in gene therapy and epigenetic engineering. 🔮 Future potential: Toward AI-driven biological design and virtual cell modeling. Evo 2 represents a major leap toward AI-powered genomic engineering. Future iterations could integrate additional biological layers—such as transcriptomics and proteomics—to create virtual cell models that simulate complex cellular behaviors. This could revolutionize drug discovery, genetic therapy, and even synthetic life creation.

Recently, a study published in Nature Immunology caught my eye. In it, the authors undertook an extensive study that charts generic variations influencing the tumour microenvironment (TME). The TME plays a crucial role in tumour progression and response to treatment. Understanding the genetic underpinnings of the TME could help pave the way for novel therapeutic approaches and enhanced treatment targeting. One of the study's most interesting aspects is its use of machine learning methods and advanced bioinformatic approaches to analyze and integrate large-scale datasets. The advanced computational methods used enabled identification of genetic variations that may have otherwise been overlooked, highlighting the power of computational biology in advancing our understanding of cancer. Leveraging these techniques, the researchers created a detailed atlas of genetic factors impacting the TME, which they refer to as immunity quantitative trait loci (immunQTLs), and showed that many of these genetic factors were likely co-localized with previously known expression quantitative trait loci. This observation suggests that the immunQTLs may contribute to the cellular heterogeneity observed within the TME by influencing the expression of genes modulating immune infiltration. Going beyond their initial discovery-driven computational work to further validate their findings, they mapped immunQTLs across >1,600 genes and 23 cancers that are associated with cancer pathogenesis and immune regulation. Diving even deeper, they went on to experimentally validate that one of the identified genes, CCL2, which is implicated in promoting colorectal carcinoma (CRC) progression by allowing tumour cells to evade immunity, may be a promising therapeutic target. This finding demonstrates the potential of the depth of the data set and how it might be used to identify and validate targets. This publication presents a significant amount of work that I have only scratched the surface of here. It offers new insights into the complexity of genetic factors influencing the TME, providing a comprehensive genetic map of the TME and its implications for cancer therapy. The authors have made their data available through a publicly accessible database to help propel further work by the research community. To me, an exciting aspect of this work is that it may help open the door to future combination therapeutic approaches that target both the tumour cells and their microenvironment. https://lnkd.in/ezRckvFh

🛠️ Builder's Notebook #2 Your grandfather's desi cow may be worth more than an imported Holstein. Let me explain why. For fifty years, India chased one number: milk yield. And on paper, the imported and crossbred cattle won more litres per day, so we bred for them, imported them, and quietly let our indigenous herds shrink. We were measuring the wrong thing. Here's what our Gir, Sahiwal and Murrah actually carry in their genetics: 🌡️ Heat tolerance. They thrive at temperatures that put exotic breeds into stress and collapse yield. In a warming world, this isn't a nice-to-have; it's survival. 🛡️ Disease resilience. Generations of natural selection in Indian conditions. Fewer inputs, fewer antibiotics, lower vet cost per litre. 🥛 Milk that commands a premium. A2 protein, higher fat and solids - the qualities global markets are now paying up for. 💪 Low-input hardiness. They convert poor fodder into value where exotic breeds need expensive, managed feeding. When you count total value per animal - not just litres, but resilience, input cost, longevity and milk quality; the maths flips. So at Verdant Genetix, we're doing what should have started decades ago: systematically genotyping our indigenous champions - Gir and Sahiwal cattle, Murrah buffalo to identify, preserve and enhance the traits that make them irreplaceable. This is genetic sovereignty. A nation that loses its indigenous germplasm is a nation that will one day rent its food security from someone else. The world is coming to us for these genetics. The question is whether Indian farmers will own that value - or watch it get exported without them. That's the future we're building at Verdant Genetix. From my village. For the 600 million. Next entry: what happens when you treat a farmer and his cattle as one health system — and why that idea became Verdant Impact Vet Labs. #BuildersNotebook #IndigenousBreeds #Genomics #GeneticSovereignty #Bharat #AgriTech

Not every shiny ML algorithm belongs in Bioinformatics. Bioinformatics doesn’t just need AI. It needs Bio-aware AI. In the rush to apply the latest AI/ML models to every problem, there’s a reality check many overlook: 👉 Bioinformatics ≠ generic tabular data. 👉 Bioinformatics ≠ simple image recognition. 👉 Bioinformatics ≠ “just another dataset.” Genomics, proteomics, structural biology, and systems biology produce data with unique statistical distributions, noise profiles, and biological constraints. - Sequence data isn’t like stock market data. - Protein structures don’t behave like social network graphs. - Gene expression matrices are not regular spreadsheets. This is why some ML models that dominate in other fields (finance, NLP, recommender systems) break down in bioinformatics unless carefully adapted. In Bioinformatics, success comes when: Algorithms are tuned for biological priors. Models respect the physics & chemistry of life. Data preprocessing mirrors the complexity of biology, not just math. The best ML algorithm is not the “newest” one, it’s the one that truly understands biological data. Here are the top ML/LLM models in 2025: - AlphaGenome (June 2025): Gene regulation & variant impact from long DNA sequences - AlphaFold 3 (Launched 2024; widely adopted by 2025): Protein complex, ligand, DNA/RNA structure prediction - SonicParanoid2 (2024): Fast orthologous gene inference using ML & LMs - NuFold (2025): RNA 3D prediction using AlphaFold 2 architecture - trRosettaRNA (Recent): Transformer-based RNA tertiary structure modeling - esmGFP / ESM3-derived protein design (Published Jan 2025): AI-designed protein simulating evolutionary processes - Generative AI Models: DNABERT, DNAGPT, GENA LM: DNA sequence modeling and classification with LLMs - EMitool (2025): Explainable multi-omics integration for cancer subtyping - DeepGO-SE and TAWFN (2025): Enhanced protein function inference via embeddings and GNNs - Graph Neural Networks (GNNs) (Growing relevance by 2025): Modeling biological networks and spatial gene expression - Quantum-Inspired Algorithms: QSVM, QNN, VQE, QFT: Experimental bioinformatics acceleration via quantum algorithms - BioMaster (2025): Automated bioinformatics pipeline management with LLM agents Models like AlphaGenome or DeepGO-SE are purpose built for biology they understand sequence context, structure, or biological ontologies. AlphaGenome handles million-base pair sequences; ESM3 was trained on hundreds of billions of protein. NuFold, AlphaFold 3, and trRosettaRNA capture 3D structure; GNNs model networks and tissue spatial contexts. Tools like EMitool and BioMaster support interpretability and autonomous workflows. Quantum-inspired algorithms and LLM agents (e.g., BioMaster) point toward the next wave of bioinformatics automation and acceleration.

🧬 AI just wrote the code for living organisms and they actually work. Researchers at Arc Institute and Stanford have achieved something unprecedented: using genome language models Evo 1 and Evo 2 to generate 16 viable bacteriophage genomes from scratch the first time AI has designed complete, functional genomes that work in the real world. Think about that for a moment. Not just designing a protein. Not simulating a genome on a computer. Actually creating living viral systems with substantial evolutionary novelty that infect bacteria and replicate successfully. Here's what makes this revolutionary: In 1977, ΦX174 was the first genome ever sequenced. In 2003, it was the first genome chemically synthesized. Now in 2025, it's the template for the first AI-generated genomes. We've gone from reading DNA, to writing it, to designing it. The results? Several AI-generated phages outperformed the wild-type virus, with one variant called EVO-Φ69 being 65x more powerful than natural viruses. One even uses an evolutionarily distant DNA packaging protein that researchers wouldn't have rationally designed. The implications are staggering: → Cocktails of these generated phages rapidly overcome antibiotic-resistant bacteria → Accelerated development of phage therapies for drug-resistant infections → A blueprint for designing synthetic biological systems at genome scale → Foundation for creating useful living systems with entirely novel capabilities This isn't science fiction anymore. AI models trained on 2 million bacteriophage genomes can now propose new genetic codes—and 16 out of 302 designs actually worked MIT Technology Review. We're watching the birth of generative biology in real-time. The ability to design life at the genomic level opens possibilities we're only beginning to imagine. What do you believe the opportunities and risks are of this virus making AI? The conversation is just beginning. 📄 Study: https://lnkd.in/ddP3Fjdp #SyntheticBiology #AI #GenerativeAI #Genomics #Biotechnology #Innovation #Science

UMBILICAL CORD DNA MAY PREDICT FUTURE METABOLIC HEALTH RISKS DNA changes in umbilical cord blood may help predict which children are at higher risk for future health issues like diabetes, liver disease, and stroke. Researchers analyzed chemical tags on DNA, known as methylation patterns, and linked specific alterations to markers of metabolic dysfunction later in childhood. Notably, changes in genes like TNS3, GNAS, and CSMD1 were tied to liver fat accumulation, high blood pressure, and abnormal waist-to-hip ratios. The study suggests that environmental factors during pregnancy may influence these early epigenetic signals. 3 Key Facts: 1. Early Risk Markers: Changes in DNA methylation at birth were linked to metabolic issues years later. 2. Gene Connections: Alterations in genes like TNS3 and GNAS were tied to liver fat and blood pressure problems. 3. Preventive Potential: Early detection could lead to proactive interventions before disease develops. Source: https://lnkd.in/gYkwtfvq

Just finished reading Google DeepMind's AlphaGenome pre-print - impressive research that builds on their AlphaFold. If you recall, #AlphaFold uses AI to predict protein structures from amino acid sequences. #AlphaGenome is its complementary counterpart, tackling an equally complex challenge: predicting DNA regulation - essentially when and where genes get turned on or off. Previous tools faced a fundamental trade-off: analyze large DNA regions but miss fine details, OR examine precise details but miss the bigger picture. AlphaGenome solves this by processing 1-million-letter DNA sequences while maintaining single-letter precision - like having both satellite and ground-level weather monitoring simultaneously. The results are impressive: 0.8-0.85 correlation with experimental data (vs. ~0.7 for previous methods) and outperforming specialized models on 22 of 24 tasks while predicting 11 different biological processes simultaneously. While it’s probably accurate to call this solid engineering progress rather than a revolutionary breakthrough (gene regulation remains as chaotic as weather prediction) it gives researchers better tools to move from statistical correlations in genetic studies to more testable biological hypotheses. For drug discovery and understanding genetic disease, this could accelerate the path from "we found a genetic association" to "here's how it actually works biologically. Once again, I am impressed by Demis Hassabis and the crew at DeepMind's contributions to advancing medical research. Here is the full pre-print: https://lnkd.in/eYuuibQz