Five short lessons follow a single human IGH sequence as it's recombined, mutated, prepared, sequenced, and observed. Every concept ends with code you can run and an exercise to make it stick.
Build a heavy-chain rearrangement segment by segment. Why NP regions exist, what trimming does, and what your first Experiment.run() actually returns.
Drag through the three pipeline phases — recombine, mutate, corrupt — and watch your sequence transform. Build a mental model of which ops belong where and why order matters.
A heatmap of where mutations actually land. Why CDRs hyper-mutate, why context matters, and how to tune rate() to match an empirical SHM range.
Live sliders for 5'/3' loss, indels, and N-bases. See exactly how each artifact distorts the molecule and which ground-truth fields stay correct anyway.
Side-by-side comparison: aligner output on real data vs. GenAIRR output on the same molecule. The case for benchmarking against truth.