NBCC Project LEAD Institute 2026, Day 2: The Same Molecules, From the Other Side

Day 1 gave me the shape of the problem. Cancer is a double coincidence — accelerator jammed on, brakes cut, same cell, accumulated over time.

Day 2 went underneath it: how information actually moves from DNA to RNA to protein, and what happens when the instructions are wrong.

I expected that to be the hard part of the week. It wasn't. The hard part was recognizing the material.

Here is what stayed with me from Day 2.

The morning was the structure and function of DNA and RNA, taught by Jaime Fornetti, PhD.

The framing that made it stick is a cooking analogy, and I'm not going to improve on it.

DNA is your grandmother's cookbook, and it can't leave her house. RNA is the copy of the one recipe you need today, and the copy can leave — out of the nucleus, into the rest of the cell. Protein is the thing you actually cooked.

So: DNA is what can be made. RNA is what you want to make right now. Protein is what you made.

Every cell in your body carries the same cookbook. What makes a breast cell a breast cell is not different DNA. It's which recipes get copied, and when — which means the control layer, not the code itself, is where most of the interesting things happen.

DNA gets damaged constantly, from the outside by UV light and X-rays and chemicals, and from the inside by ordinary copying errors. So the cell proofreads. Proteins scan for damage, checkpoint proteins stop the line, the bad section gets cut out and replaced.

Here's the sentence that reorganized the morning for me:

Normal cells wait for the repair, or they die. Cancer cells keep replicating and pass the damage to the daughter cell.

And once damaged DNA has been copied into a new cell, it is the template. It isn't damage anymore. It's the recipe.

I had been using "damage" and "mutation" interchangeably for years. Damage is an event. A mutation is damage that survived long enough to be copied.

That distinction is also the whole logic of a PARP inhibitor, which I'd never understood before. Single-strand breaks are repaired by PARP; double-strand breaks are repaired by the BRCA proteins. A PARP inhibitor blocks the first repair on purpose, so a break escalates to the kind only BRCA can fix. In a cell whose BRCA proteins are already broken, nothing can fix it, and the cell dies. The drug isn't attacking the cancer. It's opening a gap the cancer can't close and a healthy cell can.

The afternoon belonged to Ayesha Shajahan-Haq, PhD, on proteins, translation, and mutations.

She taught mutation types with a sentence. Read it in three-letter chunks, the way the cell reads codons:

The fat cat ate the wee rat

• Substitution: The fat h at ate the wee rat. One letter swapped. Still reads.
• Deletion: The fat ate the wee rat. Something missing, the rest survives.
• Frameshift: The fat ca a tat eth ewe era. One extra letter, and every word after it is garbage.

A mutation is a typo. What it costs depends entirely on where it lands and whether it shifts everything that comes after it.

There's some mercy built into the code, too — most amino acids can be spelled several ways, so plenty of single-letter errors land on the same word anyway. Some mistakes are simply allowed.

The category that stopped me was polymorphisms. A polymorphism is common, naturally occurring variation, present in at least 1% of the population. It doesn't cause cancer. But it can affect how well you detoxify carcinogens, how well your DNA repair machinery works, and how you metabolize the drugs meant to treat you.

I was diagnosed at 37 with no family history and no inherited mutation, and what I was handed was a shrug toward "environmental factors." Polymorphisms are the first thing anyone has described to me that turns that shrug into something with moving parts. Not a cause. A modifier — of how efficiently a particular body clears a particular exposure, and how well it repairs what gets through.

I have spent years thinking about the exposure side of that equation. It had not occurred to me that the repair side varies too, person to person, and that nobody ever told me which side of that distribution I was standing on.

Then she started teaching amino acids, and I had the strangest hour of the week so far.

I know these molecules. I have spent years with them. An amino acid has a carboxyl group at one end and an amino group at the other, and they link like a hook and eye — and that hook-and-eye link is a peptide bond. A short chain of them is a peptide. A long chain, folded into a three-dimensional shape, is a protein.

This is the vocabulary of my day job. Peptides are on my ingredient decks. I have written about signal peptides, about chain length, about why a molecule's shape determines whether it can do anything at all. In a formulation meeting, a peptide is an ingredient with a claim attached to it.

In that theater, the same chemistry was building the estrogen receptor. And the progesterone receptor. Which is to say: it was building the specific proteins that my own tumor was covered in, because my cancer was ER/PR+.

I want to be careful here, because these are two genuinely different contexts and I'm not going to pretend otherwise. But sitting in that room, watching a ribosome diagram click amino acids together one peptide bond at a time, and understanding that the output of that assembly line was the receptor my disease used to grow — that was not an academic experience.

It was the same science I work in every day, rotated about ninety degrees, and it looked completely different from the new angle.

The signaling vocabulary did the same thing to me. Cells talk chemically, and the categories have names: endocrine is a hormone traveling a distance, paracrine is a nearby cell sending a message, autocrine is a cell signaling itself — including telling itself to die. Estrogen is a ligand. The cells sending it are the ovaries and fat tissue.

Skincare borrows that language constantly. We say "signaling" about peptides all the time. I don't think I fully understood, until yesterday, what the word was borrowed from.

The last thing that stayed with me was about resistance, and it's the part I'll be chewing on for a while.

The textbook success story is Gleevec in chronic myelogenous leukemia. Two chromosomes swap material, producing a fused gene that makes a protein stuck permanently in the "grow" position. Gleevec fits into a pocket on that protein and shuts it off. It worked because there was one causative mutation, one simple test to find it, and one drug to block it.

Then the pocket changes shape. The drug can't bind. Growth resumes.

Someone asked how long a targeted therapy works before that happens. The answer: enormously variable. Some patients develop resistance in three months. Some go twenty years. Same drug, same disease. Nobody can currently say why.

The estrogen receptor tells a version of the same story. Mutations cluster in the ligand-binding domain — the pocket where estrogen docks. They're rare in untreated primary ER+ disease, and common in treatment-resistant metastatic ER+ disease.

The receptor changes shape under the pressure of the treatment. Which is the molecular version of the fact that stopped me on Day 1: metastases can be biologically distinct from the primary tumor that produced them.

Three things.

First, damage is not the mutation. Unrepaired damage that gets copied is the mutation — and the repair machinery is itself made of proteins that can be broken by the same forces they're defending against.

Second, "where does the typo land" is now the question I'll bring to any paper. It's the difference between a finding that matters and a finding that's decoration.

Third, resistance isn't the drug failing. It's evolution doing another draft. Three months or twenty years, same molecule, and no one can predict which — that gap is not a footnote. It's an unanswered question sitting in plain sight, and it's exactly the kind that doesn't get funded incrementally.

Day 3 is model design.