Show HN: Automating a manual step in RNA design (With a Conflicted Claude Opus)There's a huge disconnect in many life sciences labs. There are computational teams building complex models and wet lab scientists doing research at the bench. These two worlds don't always mix. This means the scientists who run experiments often get stuck with tedious, manual digital work, tasks that feel like they should have been automated a decade ago. They rely on machines for data but often lack the programming background to build the custom tools they desperately need. A concrete example from mRNA therapeutics is selecting the right gene variant for drug screening. Before a new drug can be synthesized, a researcher must choose the correct variant of a target gene (isoforms). TP53 for example has 12. Picking the wrong one can mean targeting the wrong tissue or expressing a non-functional protein, leading to false negatives that can kill a promising therapeutic before it even gets a shot. The manual workflow followed to get relevant data for any target follows a process I imagine