What We Hope The Year Will Be Like

First day, and he did not hand out a syllabus and read it to us. He put a case on the board, asked what we would need to know to solve it, and let us list things for ten minutes. Then he said the list we just made was the course.

The expectations are short enough to remember. Come in ready, write it down, say where your evidence stops.

That last one is going to come up constantly.

We walked the site: where the day's lesson lives, where the glossary is, where the folder for each kind of work is.

Everything I turn in this semester goes to one of a small number of folders, and now I know which is which.

The part I did not expect was the practice section. It is not extra, it is where the WebXam review lives, and it is open from day two.

Record, Reduce, Review, Reflect, Recite, Revise. We set up the notebook the way it will stay all semester: notes on the right, questions on the left, summary at the bottom.

He made us do the Reduce step on notes we had just taken, which was uncomfortable, because it showed how much of what I wrote down I could not actually say back.

Safety first, and not as a formality. Personal protective equipment, the Safety Data Sheet, and finding the hazard and first aid sections on a real chemical before anyone touched anything. He calls it earning entry to the lab, and that framing works on me.

Then myPLTW logins and a practice submission, so the first real one is not also the first time I have used the system. The first real lab is a week out and it is a gel, so the practice matters.

The pretest is a baseline, not a grade, and he said that before he handed it out and again after. Mine was low and so was everybody's, which was the point.

Then we wrote down a target and one specific habit to get there. Mine is fifteen minutes a night on the practice section. Seventeen weeks to move it.

Straight into safety. Personal protective equipment, then the Safety Data Sheet, and we had to find the hazard and first aid sections on a real chemical before we were allowed near anything.

He framed it as earning entry to the lab rather than a rule to obey, which landed better than I expected.

Notebook setup and the portfolio system. Everything I make this year lands in one place with a date and a label, or it does not count.

The portfolio is the thing that gets audited later. He said that twice.

AT HOME, THE NIGHT BEFORE WED SEP 2 Intervention inventory We listed every kind of medical intervention we could think of, then sorted them: prevent, diagnose, treat, replace. That four-way split is the shape of the whole course.

I had almost nothing under prevent, which turns out to be where most of the actual health impact lives.

AT HOME, THE NIGHT BEFORE WED SEP 2 Launch portfolio submission First portfolio submission plus a WebXam pretest, cold, in week one. I did badly and so did everybody, which was the point. It is a baseline, not a grade.

ETHICS DAY When can you confine someone who has done nothing wrong? Quarantine is the state deciding where a person sleeps because of a risk they carry.

I argued for public health and got asked how long is too long, and what happens when the test that justified it turns out to be wrong. I did not have a limit ready and a position without a limit is not really a position.

A sign is what someone else can observe. A symptom is what only the patient can report. Fever is a sign, nausea is a symptom.

That distinction runs the whole diagnostic process, because signs can be measured by anyone and symptoms depend entirely on the patient being believed.

Bacteria, viruses, fungi, parasites, and what actually separates them. The practical difference is that an antibiotic works on one of those four and does nothing to the rest.

Mapping who contacted whom, when. Our first map was a spaghetti mess until someone suggested ordering it by date instead of by person. Same data, completely different picture, and a cluster appeared.

First real CER of the year. Claim, evidence, reasoning. Mine had a reasoning sentence that just restated the claim in new words, which he says is the most common mistake and I have now made it.

LAB The whole class in the computer lab running BLAST, which compares an unknown sequence against a global database and ranks what it matches.

Watching an unknown resolve into a named organism in about forty seconds is the first time this course felt like the actual job. The database is public. Anyone can do this.

The part that separates a real match from a coincidence. A low E-value means the match is unlikely to be chance, and query coverage tells you how much of your sequence actually lined up.

You can get a confident-looking top hit that covers a tiny fraction of your sequence and means nothing. I would absolutely have been fooled by that yesterday.

Full report: sequence, BLAST result, E-value, coverage, and a statement of what the identification does not prove.

ETHICS DAY When a test is scarce, who is tested first? Symptomatic people, exposed people, essential workers, or whoever asks.

Testing the sickest finds the most disease. Testing the exposed finds it earliest. Those are different goals and the room split hard on which one a test is for.

Serial dilution: each step is the same fixed reduction, so ten of them get you very small very fast. It is the only practical way to get to the concentrations these assays need.

The arithmetic is easy and doing it correctly under time pressure with real tubes is not, which is why we did it on paper first.

AT HOME, THE NIGHT BEFORE THU SEP 17 Standard curve and lab prep The standard curve is the thing that makes a colour mean a number. Known concentrations, measured responses, a line through them, and then you read your unknown off the line.

Without it an ELISA tells you something is present. With it you know how much.

The sandwich: capture antibody, antigen, detection antibody, then something that produces colour. Specificity comes from the antibody only binding its target.

We modeled it dry before touching reagents, which meant Wednesday next week was checking rather than guessing.

Dilution calculations, standard curve, model diagram.

ETHICS MONDAY A false negative tells a sick person they are fine. A false positive tells a healthy person they are sick. Which error should a test be designed to avoid?

It depends entirely on what happens next. If the follow-up is a cheap confirmatory test, lean toward catching everything. If it is surgery, do not. The right answer is a property of the whole system, not the test.

AT HOME, THE NIGHT BEFORE TUE SEP 22 Controls and wet-lab plan Positive control, negative control, blank. If the negative control comes up positive, every result on that plate is meaningless.

Planning the plate layout in advance, well by well, is most of the work.

LAB Actual plate, actual pipettes, actual colour change. My hands were not steady and my volumes were not consistent for the first row, which you can see in the data.

A group near us had a negative control turn positive and had to discard the whole plate. He was almost pleased about it, because they caught it. The dangerous version is the group that never runs a control.

Sensitivity is the share of people who have it that the test catches. Specificity is the share of people who do not have it that the test correctly clears. For one given test, moving the cutoff trades them against each other: loosen it and you catch more real cases and also flag more healthy people. What breaks that trade is not a different cutoff, it is a better test.

This is Monday's ethics question with numbers attached. That ordering is not an accident.

Plate layout, raw absorbance, standard curve, unknowns, and a limitations paragraph about my unsteady first row.

ETHICS DAY Should a doctor be allowed to prescribe an antibiotic they believe will not help, because the patient expects one?

Refusing costs you the patient. Prescribing costs everyone a little resistance.

It is a tragedy of the commons in a clinic room, and the person paying is not the person deciding.

AT HOME, THE NIGHT BEFORE MON SEP 28 Antibiotic mechanisms Cell wall, protein synthesis, DNA replication, membrane. Four things to break, and each class of antibiotic breaks one.

This is also why antibiotics do nothing to viruses. A virus has none of those four targets. It has no cell wall, no ribosomes of its own, no metabolism to poison, and it copies itself inside our own cells using our machinery.

LAB Disks of antibiotic on a lawn of bacteria, and you measure the clear zone where nothing grew. For any one antibiotic a bigger zone means this organism is more susceptible to it. You cannot rank two different antibiotics by comparing zones, because each molecule spreads through agar at its own rate, so every zone is read against that drug's own breakpoint table. Getting an actual minimum inhibitory concentration takes a dilution series, not a disk.

Measuring the zone is harder than it sounds because the edge is not always sharp. Two of us measured the same plate and got different numbers, so we agreed on a rule and remeasured everything.

AT HOME, THE NIGHT BEFORE TUE SEP 29 Resistance and stewardship How resistance arises and spreads. It is not that individual bacteria learn, it is that the ones that already survive get to reproduce.

Selection pressure is why the length of a course matters. The old line was that stopping early breeds resistance.

For most common infections the evidence points the other way: the longer any antibiotic is in you, the longer it selects resistant bacteria in your own normal flora. Take the course you were actually prescribed, and do not assume longer is safer.

Zone measurements, comparison, a claim about which agent was most effective, and the measurement disagreement written into the limitations.

ETHICS DAY Should research that could produce a dangerous organism be done at all, if the same knowledge protects us later?

The honest answer nobody liked is that you cannot know in advance which it will be.

AT HOME, THE NIGHT BEFORE THU OCT 1 Aseptic technique Everything is contaminated unless you actively prevented it. Flame the loop, do not put the lid down, work near the flame, move deliberately.

The whole technique is a set of habits that assume the air is full of things trying to get into your plate. Which it is.

LAB Streak plates and colony counts. My plate had contamination in one quadrant, which is my own technique showing up as data.

Counting colonies is tedious and the number matters, so we counted twice and averaged.

Bacteria do not only inherit resistance from a parent, they can hand genes sideways to unrelated bacteria.

Horizontal gene transfer.

That is why resistance spreads faster than ordinary evolution predicts, and why a resistance gene in one species is a problem for all of them.

ETHICS MONDAY Whether a cochlear implant should be described as a fix. Some Deaf people do not consider deafness a problem needing solving, and framing the implant as a cure carries a claim about their lives.

This was the first Monday where I realized I had been assuming the medical framing was neutral. It is not. It is a framing, and it belongs to somebody.

Reading an audiogram: frequency across, loudness down, and the shape of the curve tells you what kind of loss it is.

High frequency loss first is the common pattern, which is why consonants go before vowels and people say they can hear you but cannot understand you.

LAB Modeling how a disease moves through a population and what changes when a fraction is immune.

Watching the model stop spreading before everyone is vaccinated is the clearest possible demonstration of herd immunity. The unvaccinated are protected by the vaccinated.

The threshold depends on how contagious the disease is. More contagious means a higher fraction needed, and for the most contagious diseases the number is uncomfortably close to everyone.

It is not a slogan, it is a calculation.

AT HOME, THE WEEKEND BEFORE MON OCT 12 Unit 1 tracker check Portfolio audit against the rubric. Two items were missing labels. Two minutes to fix once found.

ETHICS MONDAY If your genome shows a risk, do your relatives have a right to know? A result about you is partly a result about them.

Hardest Monday so far. I argued for privacy and then had to sit with the fact that under my own rule, somebody could learn they carry something serious and say nothing to a sibling who could act on it. I did not resolve it and my CER says so.

Reading a family tree for inheritance pattern. Dominant, recessive, sex-linked, each with a signature shape once you know what to look for.

Recessive skipping generations is the one that finally made pedigrees make sense to me.

LAB A small DNA difference changing a real trait, using the PTC tasting case. Most of whether you taste it comes down to three linked base positions in one gene, TAS2R38, inherited together as a set. Even then it is not clean.

Some people fall between taster and non-taster.

The size of the change against the size of the effect is the thing that gets you.

Writing to a family, not to a teacher. Accurate, and readable by someone frightened.

Much harder than the science. I wrote probability as a percentage and my partner said it read as a verdict, so I rewrote it as a frequency out of a hundred families, which is the same number and lands completely differently.

AT HOME, THE NIGHT BEFORE FRI OCT 16 MP1 tracker audit Marking period audit. Everything checked against the rubric with a confidence rating I had to be honest about.

ETHICS DAY Should patients get raw genetic results directly, without a clinician interpreting first? Access against the risk of misreading something frightening.

It is your information. It is also information that is easy to misunderstand in a way that causes real harm. Both things are true.

AT HOME, THE NIGHT BEFORE MON OCT 19 PCR and primers Amplifying a tiny amount of DNA into enough to work with, and how the primers decide which region gets copied.

The primers are the specificity. Design them badly and you amplify the wrong thing confidently.

LAB Running a gel. DNA is negatively charged so it moves toward the positive end, and smaller fragments travel further through the gel.

Loading the wells is genuinely difficult. I punctured the bottom of one well and lost that sample. You get one try per well.

Thousands of expression measurements at once instead of one gene at a time. The scale is the point and also the problem, because with that many measurements some will look significant by chance.

AT HOME, THE NIGHT BEFORE WED OCT 21 Methods quiz Quiz on the methods so far. This is where the fifteen minutes a night either paid off or did not. For me it mostly did.

LAB Straight into the lab after two days off, working real expression data. Which genes are turned up, which are turned down, compared against a control.

Expression is the layer between having a gene and it doing anything, which I had never really separated before.

Reading a heat map, where colour is intensity and the pattern is the finding. Red up, green down, clustered so similar things sit together.

I made a claim off a striking colour block that turned out to be three genes, which is not a pattern, it is a coincidence with good graphic design.

Expression table, heat map, claim, and a limitations line about multiple comparisons.

ETHICS MONDAY Editing a somatic cell affects one person. Editing the germline affects everyone descended from them, none of whom can consent.

That distinction does almost all the ethical work and I had not known it existed two hours earlier.

Using a virus to deliver a working gene, since delivery is what viruses are already excellent at.

The engineering problem is getting it to the right cells and not triggering an immune response that destroys the therapy.

How CRISPR targets a specific sequence, then screening embryos, which is a different thing from editing and gets confused with it constantly.

Selection and modification are not the same act and the ethics of each are different.

A full CER on where the line sits. Mine argued somatic yes, germline not yet, and had to define not yet in a way that was not just avoidance.

ETHICS DAY Who gets offered screening, and what happens when the test exists but access does not. A test nobody can reach is not a benefit, it is a statistic about a benefit.

AT HOME, THE NIGHT BEFORE WED NOV 4 Molecule-to-patient case packet Building one case from the molecular level up to the person. Sequence, expression, test result, diagnosis, intervention.

This is the first time all the units have been in one document. I had to reread my own work from September.

Sorting our evidence by how much it can carry, and marking which results could be false and what that would do to the conclusion.

A conclusion that collapses if one result is wrong is a fragile conclusion, and you should know that before you commit to it.

Recommending an actual intervention, defended with evidence from multiple units, with the uncertainty stated.

Peer review caught that I cited a test result without saying how confident the test itself is.

AT HOME, THE NIGHT BEFORE FRI NOV 6 Unit 2 summative The unit assessment. Nothing on it was a surprise, which is the whole argument for the nightly review.

ETHICS DAY Screening finds cancers early. It also finds cancers that would never have caused harm, and those people get treated anyway.

Overdiagnosis was a new idea to me. I had assumed finding more was strictly better.

LAB Establishing what normal tissue looks like before looking at anything abnormal. You cannot recognize disordered until you know ordered.

Normal cells sit in a pattern. That is the whole baseline.

ETHICS DAY Should a dying patient get an unproven drug? Compassionate access against the fact that if everyone gets it outside the trial, the trial never finishes and nobody learns whether it works.

The cruel part is that both positions are trying to save lives.

Staging is not how bad it looks, it is a defined system: size, nodes, spread. Everyone means the same thing by stage two, which is the entire point.

LAB Chemotherapy travels everywhere and leans on cancer cells dividing more often than most normal cells, which is why its side effects land on the other fast-dividing tissues: gut lining, marrow, hair. Radiation also damages DNA in dividing cells, but most of its selectivity is that you aim the beam. That is why chemo side effects are body- wide and radiation side effects are wherever the beam went.

It is a blunt instrument used carefully, not a precise one.

Drugs aimed at something specific to the tumour rather than at division in general. Fewer side effects, but only if the tumour actually has the target, which is why you test first.

This is where the genetics from Unit 2 comes back and pays off.

A recommendation for a specific patient, with the staging, the molecular result, and the uncertainty.

ETHICS DAY Who gets the organ. Sickest first, best match, longest waiting, most likely to survive.

Third allocation debate this course and my position has moved every time. There is no ordering without a bad case attached.

AT HOME, THE NIGHT BEFORE THU NOV 19 Recombinant DNA flow Putting a human gene into bacteria and having them manufacture the protein. That is where insulin comes from now.

The idea that you can hand a gene to an organism that has never had it and it just reads it is genuinely strange when you stop to think about it.

LAB Rotating stations, each with a protocol to follow exactly and data to record. Following someone else's protocol precisely is its own skill.

Growing tissue rather than transplanting it. Scaffold, cells, signals, and the vascular problem that limits how thick anything can be.

Nothing works past a couple of millimetres without a blood supply, and that is the wall the whole field is up against.

AT HOME, THE WEEKEND BEFORE MON NOV 23 Final tracker evidence packet Portfolio audit before the break.

ETHICS MONDAY Who owns an engineered protein, and what a patent on a therapy does to its price.

The argument for patents is that development is expensive and somebody has to fund it. The argument against is standing in a pharmacy.

AT HOME, THE NIGHT BEFORE TUE NOV 24 Plasmids and enzymes Plasmids as the vehicle, restriction enzymes as the scissors, ligase as the glue. Each enzyme cuts one specific sequence, which is what makes the cut predictable.

LAB The whole workflow end to end: cut, insert, transform, select, express, purify. Every step has a way to fail and a way to check.

Getting bacteria to actually make the protein, and confirming they did. Present is not the same as correctly folded and functional.

AT HOME, THE WEEKEND BEFORE TUE DEC 1 Cloning workflow quiz Quiz, then straight on with the workflow still fresh.

Checkpoint. He asked for fifteen minutes a few times this week on the unanswered pile rather than new work, which is the most reasonable homework I have ever been given.

ETHICS DAY The problem is separating one protein from thousands of others in the same tube, using whatever makes yours different: size, charge, or what it sticks to.

AT HOME, THE NIGHT BEFORE WED DEC 2 GFP and chromatography Green fluorescent protein, which is the perfect teaching molecule because you can literally see where it is.

Purification stops being abstract when the fraction you want glows.

LAB Running the column and collecting fractions. Watching the glow move down the column and come out in a specific set of tubes is the most satisfying thing this course has done.

I collected one fraction late and lost part of the peak, which shows up in the yield.

Running the fractions on a gel to see what is actually in each one. Purity is visible as the other bands disappearing.

My best fraction still had two faint contaminant bands, which is honest and went in the report.

Fractions, gel image, yield calculation, and the late-collection error written into the limitations.

ETHICS MONDAY The allocation question again, now with the actual matching criteria in front of us instead of in the abstract.

Having the real criteria made my position sharper and less comfortable.

AT HOME, THE NIGHT BEFORE TUE DEC 8 Nephron and renal data The nephron in order: filtration, reabsorption, secretion. The kidney filters nearly everything out and deliberately takes most of it back, which is how it keeps such precise control.

LAB Dialysis as an external substitute for filtration, then a matching simulation where you allocate organs under the real constraints.

Doing it as a simulation, with actual patients on cards, is completely different from arguing about it on a Monday. Somebody does not get one.

Blood type, tissue type, and why the immune system attacks a transplant that is helping the person it belongs to.

Immunosuppression is a permanent trade.

A CER defending an allocation decision from the simulation, using the compatibility data.

ETHICS DAY How much testing before a new technology reaches patients. Speed against safety, and the people harmed by delay are invisible in a way the people harmed by a bad device are not.

AT HOME, THE NIGHT BEFORE MON DEC 14 Scaffolds and stem cells Scaffold for shape, cells to populate it, signals to tell them what to become. Then the different stem cell types and what each can and cannot do.

LAB Animal organs and mechanical replacement as two different answers to the same shortage, with different failure modes and different ethics.

Analyzing a real device against what the patient actually needs rather than what is technically impressive.

A written critique naming one strength, one weakness, and one thing that would have to be true for it to work.

ETHICS DAY The last one. We were asked to name one position we had changed across the whole course.

Mine was genetic privacy. I came in certain and I am leaving genuinely unsure, and I now think that is an improvement rather than a failure to decide.

One case carried from molecule to patient to intervention, using evidence from every unit.

I pulled from the BLAST identification, the ELISA, the expression data, and the treatment memo. Four units, four months apart, one argument.

IN CLASS, FRI DEC 18 Portfolio audit Every artifact checked against the rubric. Mine was mostly complete and the gaps were all labels and dates rather than missing work, which is its own lesson.

IN CLASS, FRI DEC 18 Gap closeout Fixing what the audit found. Two hours of small corrections that would have been ten minutes if I had done them at the time.

IN CLASS, FRI DEC 18 Final portfolio submit Submitted. Seventeen weeks, four units, one portfolio.

In August I could not tell a sign from a symptom. Today I built a case from a DNA sequence to a treatment recommendation and defended every step with my own data.