Manuel Gentiluomo

The human genome contains 3 billion bases. Reading them all, in the right order, is one of the greatest challenges in modern biology. For decades, Illumina ruled genomics: precise, cheap, scalable. But with one major blind spot it reads fragments of just ~150 bases at a time.

💡 Slide 4 — Why it matters: long reads span repetitive regions invisible to Illumina, detect structural variants, and read epigenetic modifications directly, no extra chemistry required.

The human genome contains 3 billion bases. Reading them all, in the right order, is one of the greatest challenges in modern biology. For decades, Illumina ruled genomics: precise, cheap, scalable. But with one major blind spot it reads fragments of just ~150 bases at a time.

🔬 Slide 2 — PacBio SMRT: a single polymerase sits inside a nanoscale well (ZMW). Fluorescent nucleotides flash one by one as they are incorporated.

The human genome contains 3 billion bases. Reading them all, in the right order, is one of the greatest challenges in modern biology. For decades, Illumina ruled genomics: precise, cheap, scalable. But with one major blind spot it reads fragments of just ~150 bases at a time.

Like trying to understand a novel by reading three words at a time. Then came third-generation sequencing. Nanopore and PacBio read thousands or millions of bases in a single read. In real time. No amplification needed. How does it work? Swipe through the slides 👆

From AlphaGo’s “Move 37” to AlphaFold, and now systems like AlphaProof and AlphaEvolve: in just a decade, AI has moved from defeating human champions to contributing to science, mathematics, and code optimisation. #AlphaGo #AI

2/ This isn’t just a story of better performance — it’s a paradigm shift. We’re seeing models that don’t just execute tasks, but explore, plan, and discover new solutions in complex spaces. #DeepMind #AGI