When they start operating you will see the patient here.
对方开始手术后，你就会在这里看到病人。
The patient will talk to you here.
🩺 You are the other doctor
You are not touching the DNA — so you are the one the patient trusts. Listen to what
they ask you, and answer them out loud in English. You will need all of it for
the argument at the end.
你没有动他的 DNA，所以他信任你。听他问什么，用英语大声回答。
🛸 Earth, the year 2050
Listen carefully. Your visitors are speaking.
仔细听。访客正在说话。
Before you start — say this to your partner:
“Hello, I am Gene Doctor One. What is your name?”
“I think the biggest problem on Earth is ______.”
“Are you ready to open the DNA?”
先和同伴说这几句，再开始。
👥 Two doctors, one laboratory
This game needs two students and one microphone.
这个游戏需要两名学生和一个麦克风。
Doctor 1
Talks first in the discussion. Modifies a human alone, then tells Doctor 2 what happened.
Doctor 2
Answers in the discussion. Modifies a different human alone, then tells Doctor 1 what happened.
👥 Who is playing?
Answer this first — it decides whether there is a second doctor at all.
先回答这个：它决定了这一局有没有第二位医生。
Solo: you choose one human, run one operation, and see one case file.
Doctor 2 is switched off entirely.
单人模式：只选一个人、做一次手术。第二位医生的部分完全不会运行。
📡 Second screen — room—not connected
The doctor who is not operating opens this same page on their own phone, taps
👁 Watch, and types the code. They will see the patient change live — and the
patient will talk to them.
不动手的那位医生用自己的手机打开同一个页面，点“👁 Watch”，输入房间号，
就能实时看到病人的变化——病人也会跟他说话。
🎬 The Reality of DNA — 9 minutes, English + 中文 subtitles
Watch this first. It explains what DNA actually is — not just what it
is like. 先看这个视频，它讲的是 DNA 到底是什么，而不只是像什么。
🔬 How genes work
Five machines to play with. Touch everything, and say what you see.
五台机器，随便玩。看到什么就说出来。
💪 What is the strongest force in nature? the strong nuclear force
It has a name: the strong nuclear force, one of the four fundamental forces (with
gravity, the electric force you meet in the base-pair lesson, and the weak force). It is by
far the strongest — about 100 times stronger than the electric force, and roughly
10³⁸ times stronger than gravity. But it has an incredibly short reach: it acts only
across about the width of a proton, then vanishes.
It is carried by particles called gluons — literally the “glue.” Quarks throw gluons back and
forth, and that constant exchange is the grip. The strange part: unlike every other force, the
strong force gets stronger the further apart the quarks move, like a stretched rubber band.
Pull hard enough and the band's energy just makes new quarks rather than letting one escape —
which is why a lone quark has never been seen. Quarks are trapped for good; physicists call
this confinement.
It also does a second job. A little of it spills out past each proton and neutron and glues
them to one another — that leftover pull is what holds the nucleus together, overpowering the
electric repulsion of all those positive protons that would otherwise blow it apart.
And here is the astonishing part: the three quarks themselves weigh almost nothing — only about
1% of the proton's mass. The other 99% is the energy of the strong force,
turned into mass by Einstein's E = mc². So nearly all of your body's weight is not “stuff” at
all — it is bound-up energy. Release a tiny bit of it and you get the Sun and
nuclear power.
“A, T, G and C are letters” is not true. They are four molecules.
Their real names are long — adenine, thymine, guanine, cytosine — so we write
them by their first letter, the same way we write km instead of
kilometre. The letter is a short name, not the thing itself.
A、T、G、C 不是字母，是四种分子。它们的真名很长，所以用第一个字母来代替。
Click a base
🧬 Now build one. Snap the right partner onto each rung — a big base always
faces a small one, so A only takes T and G only takes C.
Pair every rung and your flat ladder winds into a double helix — a spiral
staircase that turns once every ten steps.
现在自己搭一个。把正确的碱基扣到每一级上——大的总是配小的，所以 A 只配 T，G 只配 C。把每一级都配好，平梯子就会拧成双螺旋——像旋转楼梯，每十级转一整圈。
Why can A only hold T?
1. Width. The ladder must stay exactly the same width the whole way up —
about 2 nanometres. A and G are purines: two rings joined
together, the big ones. T and C are pyrimidines: one ring,
the small ones. A big one must always face a small one. Two big ones would
bulge; two small ones would pinch.
2. The bonds must line up. “Holding hands” means hydrogen bonds —
a weak pull between a hydrogen atom on one base and a nitrogen or oxygen atom
on the other. They only line up when the shapes match. A and T line up in
two places. G and C line up in three. That is why G–C is the
stronger step.
Not bases. The two long sides are a chain of alternating sugar and
phosphate. The sugar is called deoxyribose — that is the
D in DNA. Each base hangs off a sugar and points inward,
so the code sits protected inside the twist.
One full step is: sugar–phosphate | base ··· hydrogen bonds ··· base | phosphate–sugar
梯子的两边不是碱基，而是糖和磷酸交替连成的链。
糖叫脱氧核糖，就是 DNA 里的 D。碱基朝内，保护密码。
What is a “ring”?
nitrogen
carbon
shared edge = fused
A ring is a closed loop of atoms bonded to each other — a real circuit,
not a way of drawing it. In these four molecules the loops are built from
carbon and nitrogen atoms. That nitrogen is why A, T, G and C are
called nitrogenous bases.
One ring (T and C) is six atoms: four carbon, two nitrogen. Two rings (A and G) is a six-atom ring fused to a five-atom ring —
fused meaning they share two atoms, so it is one solid piece, not two
rings tied together. Nine atoms in total.
Two things follow, and they are the whole reason the pairing rule works: 1. Two fused rings are simply wider than one. That is where “big”
and “small” come from — it is not a label, it is a measurement. 2. The rings are flat and stiff. The electrons are shared right
around the loop, which locks every atom into one plane. You cannot bend, fold
or twist a base. So the shape rules are exact, not approximate.
Not a normal bond. A normal bond (the ones inside the rings) shares
electrons and is strong. A hydrogen bond only borrows, and it is
about twenty times weaker.
It works like this. When a hydrogen atom is bonded to a nitrogen or an
oxygen, that bigger atom pulls the shared electrons towards itself. The
hydrogen is left slightly positive; the N or O is left slightly
negative. A hydrogen bond is the pull between that slightly positive
hydrogen and a slightly negative N or O on the other base. Opposite charges
attract — that is all it is.
But why does the big atom win the electrons? Because nitrogen and oxygen
are greedy for electrons — chemists call that pull electronegativity. Hydrogen is
generous; N and O are bullies. The shared electrons drift towards the N or O, leaving the
hydrogen almost bare — and a bare hydrogen is just a proton: a tiny, naked +. That
concentrated little charge is what makes hydrogen such a good gripper for its size.
A hydrogen bond is also fussy about direction — it is strongest when the three atoms sit
in a nearly straight line, N–H ··· O. Because it has to line up so precisely, the two
bases must be the exact right shape to meet. That is the real reason A fits only T,
and G only C. One base gives the hydrogen (its N–H or O–H); the other offers a
lonely N or O to catch it.
So “holding hands” is really a weak electrical pull across a small gap.
And weak is exactly what DNA needs: the bonds must break every time a cell
copies itself, so the ladder can unzip. Strong bonds would make the code
permanent and useless.
One is feeble; two or three side by side, repeated over millions of steps, hold
the whole helix shut. Water does the same trick — that is why ice floats
and why water is sticky.
🧱 And they help hold you together. One hydrogen bond is weak, but life
uses them by the billion. They fold every protein into its exact working shape and
act as the glue in your body's structural stuff — the collagen in skin and tendons,
the keratin in hair and nails — and in wood, cotton and spider silk. Cook an egg and
you break its hydrogen bonds: the runny white unfolds and sets solid. The same faint
pull that zips DNA gives living things much of their shape and solidity.
🧱 氢键也把你连在一起。一个氢键很弱，但生命用上亿个。
它们把每个蛋白质折成精确的形状，是身体结构物质的“胶水”——皮肤和肌腱里的胶原蛋白、
头发和指甲里的角蛋白，还有木头、棉花和蛛丝。煮鸡蛋就是打断氢键：透明的蛋清展开、变硬。
把 DNA 拉在一起的那点微弱吸引力，也给了生命大部分的形状和坚实。
氢键不是普通的化学键，它不共用电子，只是“吸引”，弱大约二十倍。
氢连着氮或氧时，氢带一点正电，氮氧带一点负电；氢键就是这一点正电和另一个碱基上一点负电之间的吸引力。
为什么大原子会赢？因为氮和氧“贪电子”（叫电负性），氢很大方；共用电子被拉向氮氧，氢几乎只剩一个裸露的质子——一个很小的正电荷。
氢键还很挑方向：三个原子几乎排成一条直线（N–H···O）时最强。正因为必须对得这么准，两个碱基的形状要正好匹配——
这才是 A 只配 T、G 只配 C 的真正原因。弱是必要的：细胞复制时梯子必须能拉开。水分子之间也是氢键。
🧲 Feel it — why + and − pull together
Everything is built from atoms, and atoms hold tinier bits: protons with a
positive (+) charge and electrons with a
negative (−) charge. “Charge” is just a property they carry —
there are only two kinds, and Benjamin Franklin named them + and − long ago. Equal + and −
make an atom neutral; pull electrons away and a spot turns +, pile them up and it turns −.
The rule: opposite charges attract (pull together); same charges repel (push apart)
— exactly like magnets. Each charge makes an invisible field around it, and any other
charge inside that field feels a push or a pull. Why? Because this is the
electric force — one of the basic forces of the universe, like gravity. We can predict it
perfectly, but “why” stops here: it is simply how reality is built.
A hydrogen bond is that force in miniature: a slightly-+ hydrogen
pulled to a slightly-− N or O. Flip the right charge below and
watch attract turn into repel.
All of it at once — real rings, not dots. A big 2-ring base (A/G)
faces a small 1-ring base (T/C), joined by dashed hydrogen bonds — 2 for
A–T, 3 for G–C — spiralling up the twist.
把一切放在一起：真正的环（不是点）。大的两环碱基配小的一环碱基，用氢键相连
（A–T 两个，G–C 三个），沿着螺旋上升。
🌀 The ladder is really a spiral staircase
Drawing DNA flat is a convenience, not the truth.
A real ladder of DNA is twisted — it makes one full turn every ten steps.
That is what double helix means: two strands, both spiralling, never touching.
把 DNA 画成平的只是为了方便。真正的 DNA 是拧着的，每十级台阶转一整圈——
这就是"双螺旋"的意思：两条链一起螺旋上升，永不相碰。
the two
sugar-phosphate sidesA–T stepG–C step
🤔 Why does it twist at all? Because water pushes
Nothing is twisting it. The shape is what the molecule falls into when you drop it
in water, and it comes from a disagreement.
The sides of the ladder — sugar and phosphate — love water. The bases
in the middle are flat, greasy and hate water. So the bases huddle inward, stacking
face to face like a roll of coins to hide from the water, while the sides stay outside where
the water is. But the rungs are attached to the sides at a slight angle, so a stack of
them cannot go straight up — each step is rotated about 36° from the one below.
Ten steps × 36° = 360°: one full turn. That is where "ten" comes from.
The twist leaves two dents running up the outside — a wide groove and a
narrow groove. This matters enormously: it is how a protein can read the code
without opening the ladder. It reaches into the wide groove and feels the edges of the
bases, like reading a book through a gap in the spine. Cas9 checks its guide this way, and so
does every protein that switches a gene on or off.
Two metres of DNA is folded into every cell in your body, inside a nucleus far
too small to see. To use a gene, the cell has to find it, open it, and copy it out —
without ever letting the original leave. Drag the slider, or press play.
你身体的每个细胞里都塞着两米长的 DNA，装在一个看不见的细胞核里。
要用一个基因，细胞必须找到它、打开它、把它抄出来，而原件永远不许离开细胞核。
拖动滑块，或者按播放。
The whole journey in one line: DNA → RNA → protein
The DNA never leaves the nucleus. It is the master copy — too long, too precious, and
far too easy to damage. So the cell does what you would do with a rare book in a library:
it copies out only the page it needs and carries the copy away to work from.
That copy is called RNA. The machine that reads the copy and builds the actual product
is called a ribosome, and the product is a protein — the thing that finally does
a job in your body. Every gene you have ever heard of works this way: a stretch of DNA is
opened, copied into RNA, and the RNA is used as instructions to build one protein.
DNA 永远不离开细胞核——它是母本，太长、太宝贵、太容易损坏。
所以细胞像对待图书馆的珍本一样：只抄出需要的那一页，拿抄件去干活。
这份抄件叫 RNA；读抄件、真正把产品造出来的机器叫核糖体；产品叫蛋白质——
最后真正在你身体里干活的东西。
🔡 How the instructions are actually written — the genetic code
So where is the instruction stored? Not in the shape of anything — in the order of the
four letters. A gene is a sentence spelled in an alphabet of four: A, T, G, C. Re-order
the letters and you change the message, exactly as re-ordering letters changes a word. That
order is the entire information; nothing else about the molecule carries meaning.
The cell reads that sentence in words of exactly three letters. Why three? Four letters
taken one at a time make only 4 words; two at a time, 16 — still fewer than the
20 amino acids a protein is built from. Three at a time make 4×4×4 = 64 words
— more than enough. Each three-letter word is a codon, and almost every codon names one
amino acid.
64 words for 20 amino acids means the code is redundant: most amino acids have several
codons, so many single-letter typos still land on the same amino acid — a built-in spell-check.
Three codons mean STOP (end of protein), and AUG means both “start here” and
the amino acid methionine. AUG sets the reading frame — where the three-letter words begin
— and everything is counted in threes from there. Shift the frame by one letter and every word
after it is misread (the Delete step again).
Then it is used: the gene is copied into mRNA, a ribosome walks the mRNA three letters at a
time, a matching tRNA drops off each named amino acid, and the chain folds into the protein. The
same code runs in a bacterium, a banana and you — which is why a human insulin gene still works
when placed inside yeast, and how insulin is made today.
Build a codon — pick three mRNA letters and see what it names
Watch it happen — gene → mRNA → protein.
Drag the slider or press play.
看它发生一遍——基因 → mRNA → 蛋白质。拖动滑块或按播放。
📖 Step 1 — opening the book at one page only
Nothing tears the DNA. The two strands are held together only by hydrogen bonds —
the weak pull you met in the base-pair lesson — and weak is exactly the point. A protein
called a helicase runs along the ladder and simply pulls the two halves apart,
like opening a zip. No cutting, no damage: the ladder closes again behind it.
This is why the pairing rule matters so much. G–C holds with three bonds, A–T with only
two, so a stretch rich in A and T opens more easily — and genes tend to have an
easy-to-open patch right at the start, where the machinery needs to get in.
Crucially, the cell opens one gene, not the book. The rest stays wound up and shut.
Which pages get opened is the entire difference between one kind of cell and another.
A machine called RNA polymerase lands on the opened gene and crawls along one of the two
strands, reading it letter by letter and building a matching copy as it goes — using the
same pairing rule you already know. Behind it, the ladder zips shut again.
The copy is RNA, not DNA, and it differs in three small ways: it is
single-stranded (a copy needs no partner), its sugar is slightly different, and it uses
U in place of T. So where the DNA said A, the RNA copy says U.
This particular copy has a job title: messenger RNA, or mRNA — it is a message
being carried out of the nucleus. It squeezes through a hole in the nuclear wall called a
pore and arrives in the main body of the cell. The original gene, untouched, closes
behind it.
This is also what a vaccine can be. An mRNA vaccine skips the DNA entirely and hands
your cells the message directly — which is why it can never change your genes: it never goes
near them, and the message is thrown away within days.
一台叫 RNA 聚合酶的机器落在打开的基因上，沿着其中一条链爬，
一个字母一个字母地读，同时照着配对规则造出一份对应的抄件；身后梯子重新合上。
抄件是 RNA，和 DNA 有三点不同：单链、糖略有不同、用 U 代替 T。
这份抄件叫信使 RNA（mRNA），它从核孔挤出细胞核。mRNA 疫苗就是直接把这条信息
交给你的细胞——所以它永远不可能改变你的基因：它根本不靠近基因，而且几天内就被丢弃。
🏭 Step 3 — the ribosome builds the protein
Out in the cell, a ribosome clamps onto the mRNA and pulls it through like tape through
a player. It reads the letters in groups of three. Each triplet is called a
codon, and each codon names one amino acid — there are twenty amino acids, and
they are the beads the protein is made of.
Small delivery molecules called tRNA float in carrying one amino acid each. A tRNA can
only dock if its own three letters match the codon showing in the ribosome, so the right bead
arrives at the right moment. The ribosome joins it to the chain, shifts along three letters,
and repeats — about ten codons every second.
When the chain is finished it does not stay a chain. It folds, in milliseconds, into one
exact three-dimensional shape, and that shape is the function. A protein folded into a
pocket that grips one molecule is an enzyme. Folded into a long fibre it is hair or muscle.
Cas9 is one of these shapes too.
Now the Delete step makes sense. If the repair machinery loses one letter, every group
of three after that point is misread — the beads come out in the wrong order and the shape is
gone. That is why losing a single letter can switch off a whole gene.
👁️ A real one — the protein that colours your eyes. Let us run the whole pipeline once
for something you can see in the mirror.
The colour of an eye is really one brown pigment called melanin, sitting in the
front of the iris. A lot of it makes eyes brown; a little makes them green or hazel;
almost none, and the iris simply scatters light and looks blue — blue is not a pigment at
all, it is the same trick as the sky.
Making that pigment needs a protein, and the protein has a gene: OCA2. Here is its
journey — the exact three steps above: 1. Open — in an iris cell, a helicase unzips the OCA2 gene; the rest of the book stays shut. 2. Copy — RNA polymerase copies it into mRNA, which slips out of the nucleus. 3. Build — a ribosome reads the mRNA in codons and builds the P protein, which folds
into a shape that helps pack melanin into the iris.
More working protein → more melanin → browner eyes. Turn the gene down and you get less
melanin and blue eyes. In fact almost everyone with blue eyes shares one tiny change in a
switch called HERC2 sitting next to OCA2 — a single letter that first appeared in one person
thousands of years ago and quietly dialled the gene down. Remember the Delete step: change one
letter and you change the protein — and here you can watch the result looking back at you.
🧠 Same book, different pages — why a skin cell is not a brain cell
Almost every cell in your body carries the identical full set of DNA. A skin cell has the
brain genes. A brain cell has the skin genes. Nothing was thrown away.
The difference is entirely which pages are open. In a skin cell the skin genes are
unwound and readable while the rest is packed away tight; in a brain cell it is the other way
round. Cells lock pages shut by winding that stretch hard around its protein beads and adding
chemical tags to it, and they can hold a page shut for a lifetime — and pass that setting on
when they divide.
Only about 2% of your DNA is genes at all. Much of the rest is switches and volume
controls that decide which pages open, in which cell, at which moment. The eye-colour switch
you met in the last lesson is exactly one of these.
你身体里几乎每个细胞都带着完全相同的全套 DNA——
皮肤细胞里有大脑的基因，大脑细胞里也有皮肤的基因，什么都没丢。
差别完全在于哪几页是打开的。细胞把要关掉的那一段紧紧缠在蛋白珠子上并贴上化学标签，
可以关一辈子，还能在分裂时把这个设置传下去。
你的 DNA 里只有约 2% 是基因，其余很多是决定"哪页在哪个细胞什么时候打开"的开关。
🌍 The whole genome — your complete instruction book
Everything so far has been one gene. Your genome is the whole book: every
letter of DNA in one cell, all 46 chromosomes read end to end. It is about 3 billion
letters long — printed at normal size it would fill a stack of books taller than you, and
read aloud without stopping it would take over 90 years.
Yet only about 20,000 genes hide inside it — a genuine surprise, because a grain of
wheat has more. Genes are just about 2% of the book; the other 98% is switches,
spacers and old broken-down copies, and we are still working out what much of it does. And
any two people's genomes are about 99.9% identical — every face, every skin tone, every
difference between us lives in the last 0.1%.
Reading the first whole human genome — the Human Genome Project — took 13 years and
about three billion dollars, finishing in 2003. The same read now costs a few hundred dollars
and takes about a day. That collapse in price is what makes modern medicine — cancer
screening, rare-disease diagnosis, and the gene editing in the next lesson — possible at all.
🌱 Stem cells — cells that have not chosen a page yet
If a skin cell and a brain cell differ only by which pages are open, one question
follows: what about a cell that has not opened any page yet? That is a stem cell
— a cell that still holds every option open and can become many different kinds of cell.
The ultimate stem cell is the fertilised egg: from that one cell come all 200-odd cell
types of a whole body. As an embryo grows, cells commit — lock most pages shut — and
become specialists. A few stem cells stay behind in the adult, quietly topping up blood, skin
and gut for a lifetime.
Stem cell therapy uses these blank-slate cells to replace what is lost. The
oldest one, the bone-marrow transplant, has run for over 50 years: it hands a leukaemia
patient a healthy donor's blood stem cells, which rebuild an entire new blood system. Newer
work coaxes stem cells into insulin cells for diabetes or nerve cells for spinal injury.
The Nobel-winning trick (2012) was to run differentiation backwards: add four switch
proteins and an ordinary skin cell reopens all its pages, becoming an iPS cell —
a stem cell made from a patient's own body, so no donor and no rejection. Join this to the next
lesson and you get the frontier: take a patient's own cells, CRISPR-fix the faulty gene,
grow them into what is needed, and put them back. The approved sickle-cell cure, Casgevy,
is exactly this — a patient's own blood stem cells, edited and returned.
🧬 The other job — copying all of it before a cell divides
Reading is the everyday job. There is a second, rarer one: when a cell is about to
divide, the whole three billion letters must be copied so each new cell gets a
complete set.
The double helix makes this beautifully easy, and it is the reason the structure matters.
Unzip it, and each half already carries the full message — because every letter names
its partner. A strand reading ATGC can only ever have faced TACG. So the cell
opens the ladder and builds the missing side onto both halves at once, ending with two
identical ladders where there was one. Each new ladder is half old, half new.
The copying machine proofreads as it goes and fixes almost everything it gets wrong. Almost:
roughly one letter in a billion survives as an error. Across three billion letters that
is a small handful of changes per copy — too few to break you, but never zero. Those leftovers
are mutations, and they are simultaneously the cause of cancer and the raw material of
evolution. Copy DNA perfectly and nothing could ever adapt.
This is the real job. Four tools, four verbs — the same words you
will shout in the lab. 这就是真正的工作，四个工具，四个动词。
Your task
What is actually happening here — and where is the enzyme?
You are doing real gene surgery, in four steps — the same four verbs you will shout in
the lab: cut the DNA open, delete the broken gene, insert the healthy
one, then close and repair the strand. Each verb is one move of a real molecular
machine, and the strand above shows the result of the move you just made.
So why doesn’t the enzyme show up in the ladder? Because the ladder is the
DNA — sugar-and-phosphate sides, base-pair rungs — and the enzyme is not made of DNA at
all. It is a separate protein: a tiny machine that floats in, grips the ladder at
one exact spot, snips it, and lets go. Think of scissors cutting a rope — the scissors do
the cutting, but they are not part of the rope. The real gene-editing enzyme,
Cas9, works just like that: a short guide molecule steers it to one address out of
three billion, it cuts both strands, then it leaves. That is why it never appears as a
rung — it only acts on the rungs.
你在做真正的基因手术，四个步骤：切开、删除、插入、合上。酶为什么不在梯子里？
因为梯子就是 DNA，而酶根本不是 DNA——它是一种蛋白质，一台独立的小机器，像剪刀一样切开 DNA 然后离开。
剪刀会剪绳子，但剪刀不是绳子的一部分。真正的基因剪刀 Cas9 就是这样工作的。
🔬 Open a tool — who really does the work?
The four buttons make it look like you do all four jobs. In a real cell
only the cut is yours — the other three are done by the cell's own repair
machinery, and all you can do is load the dice.
四个按钮看起来像是你在做全部四件事。在真实的细胞里，只有"切"是你做的，
其余三步都是细胞自己的修复机器在做，你只能影响它更可能走哪条路。
✂️ Cut — Cas9, the only step you control
You deliver a Cas9 protein loaded with a guide RNA about 20 letters long.
Cas9 does not read the whole genome looking for your target — it first hunts for a
short landing signal called a PAM (for the common Cas9, just the letters
NGG). No PAM nearby, no cut, however perfectly your guide matches.
On a PAM hit it unzips the helix and holds the guide against the exposed strand.
If they pair, two cutting parts fire at once: the HNH domain cuts the strand that
matches the guide, and the RuvC domain cuts the other one. The result is a clean
double-strand break about 3 letters in from the PAM.
One thing the scissors picture gets wrong: Cas9 does not spring away. It clamps onto
the cut ends for hours — long enough that which repair path the cell takes is
partly decided by how long Cas9 keeps sitting there.
🗑️ Delete — the cell panics, and the scar kills the gene
There is no delete tool. Nothing goes in and lifts the broken gene out.
A snapped chromosome is an emergency, so the cell throws its fastest repair crew at it:
non-homologous end joining. A clamp called Ku grabs both loose ends, other
proteins chew the ragged edges tidy, and the two ends are welded straight back together.
It is fast but sloppy — it usually loses or gains a few letters at the join.
That tiny slip is the whole trick. DNA is read in groups of three, so adding or
removing one or two letters shifts every group after the cut and the rest of the recipe
becomes nonsense. The gene is now dead. A "gene knockout" is a scar, not a removal —
the broken gene is still physically there, it just cannot be read.
To really remove a whole gene you cut twice, with two different guides, and let the
cell join the two outer ends — the piece in the middle is simply left out.
🧬 Insert — the cell copies a template you must supply
This is the hard one, and the reason gene therapy is still difficult.
The cell has a second, accurate repair path: it can rebuild a broken stretch by
copying from a matching template. Normally the template is the other chromosome.
To insert a gene, you flood the cell with your own template — your new gene with
short matching ends ("homology arms") that look exactly like the DNA either side
of the cut. If the cell picks up your template instead, it copies your new gene in as if
it were restoring a backup.
You are running a race you usually lose. This accurate path only works in cells that
are actively dividing, and the fast sloppy path from the Delete step is competing
for the same broken ends. Success rates are often only a few percent.
That imbalance shapes real medicine: most approved CRISPR treatments — such as the sickle
cell therapy Casgevy — switch a gene off rather than install a new one,
because breaking is easy and installing is not.
The last step is the quietest. A separate enzyme called DNA ligase re-forms the
sugar-and-phosphate backbone — the two long sides of the ladder — so the strand is
one continuous piece again. The cell keeps different ligases for different jobs:
LIG4 finishes the fast sloppy repair, LIG1 finishes the accurate copying one.
Note who is not here: Cas9 cannot close what it opened. It is a cutter and
nothing else. Every gene edit therefore needs at least two different proteins that never
meet — one you deliver, one the cell already owns.
And this is why the strand shows a healthy trait only at the very end. Until the backbone
is sealed, the cell treats the chromosome as damaged and will keep trying to repair
it — including in ways you did not want.
最后一步最安静：一种叫 DNA 连接酶的酶把糖—磷酸骨架重新焊好，
让链重新变成完整的一条。不同的修复用不同的连接酶：马虎修复用 LIG4，精确修复用 LIG1。
注意这里没有谁：Cas9 合不上它自己切开的口子，它只会切。所以每一次基因编辑至少需要两种
互不见面的蛋白质——一个你送进去，一个细胞本来就有。骨架没焊好之前，细胞一直把染色体当成
"受损"，会不停地想去修它，包括用你不想要的方式。
Why do some traits show and some hide? A dominant gene wins;
a recessive gene only shows when there is no dominant one. Here it is in a trait
you can see on each other's faces — brown eyes and blue eyes.
B = brown (dominant), b = blue (recessive). Change the parents and watch
the children change.
为什么有些性状显现、有些隐藏？显性基因会赢，隐性基因只有在没有显性基因时才显现。
这里用一个你能在彼此脸上看到的性状——棕眼和蓝眼。B 是棕色（显性），b 是蓝色（隐性）。
改变父母，看孩子怎么变。
Mother母亲
Father父亲
A B allele carries a working recipe for melanin — brown pigment — in the
iris. One B makes enough to fill the iris with brown, so B b still looks
brown. Only b b leaves the iris almost empty of pigment.
B 是"造黑色素"的好配方。一个 B 就够把虹膜填满棕色，
所以 B b 也是棕眼。只有 b b 的虹膜几乎没有色素。
Why does the dominant gene “win”?
You carry two copies of most genes — one from your mother, one from your father.
A gene is a recipe for a protein. A dominant allele (big B) carries a
recipe that makes a working protein — here, melanin, the brown pigment. A
recessive allele (small b) makes little or none.
So one working copy is enough. With B b, the B still fills the iris with brown,
so the eyes look brown — that is exactly what “dominant” means. The trait only
hides — blue eyes — when both copies are the quiet one, b b. That is why
blue can skip generations: two B b parents both have brown eyes, yet each can
pass on a hidden b, and a b b child has blue eyes.
Reading the square: each box is one equally-likely child. Count the boxes that contain
a big B — those children have brown eyes; only the b b box is blue.
你的每个基因有两份（母亲一份，父亲一份）。基因是蛋白质的“配方”。
显性基因（大 B）能做出有用的蛋白质——这里就是黑色素，棕色色素；隐性基因（小 b）几乎不做。
只要有一份好的就够了，所以 B b 也是棕眼——这正是“显性”的意思。只有两份都是安静的那一个（b b）时，
性状才隐藏，变成蓝眼。这就是蓝眼会隔代出现的原因：两个 B b 的父母都是棕眼，却各自可能传下一个隐藏的 b，
生出蓝眼的孩子。方格里每个格子是一个等概率的孩子：带大 B 的是棕眼，只有 b b 是蓝眼。
🌈 Blue eyes contain no blue pigment at all
Brown eyes have brown pigment in them. Blue eyes do not have blue pigment — there is
no such thing in a human iris. There are only two pigments available, a brown-black
one and a red-yellow one.
A b b iris is simply nearly empty. Light goes into the cloudy front layer,
bounces around among tiny fibres, and the short blue wavelengths scatter back out
strongest while the longer ones pass deeper and are absorbed. So the blue you see is not a
colour the eye contains — it is a colour the eye throws back at you.
This is the same physics that makes the sky blue. It also means a blue eye has no
blue to lose: what changes between people is only how much brown is in the way.
That is why babies are often born blue-eyed and darken over the first year — the melanin
has not been made yet.
棕色眼睛里真的有棕色色素；蓝色眼睛里没有蓝色色素——
人的虹膜里根本不存在蓝色色素，只有一种棕黑色和一种红黄色。b b 的虹膜几乎是空的：
光进入前层，在细小纤维之间散射，波长短的蓝光被散射回来最多，长波则被吸收。
所以蓝不是眼睛"含有"的颜色，而是眼睛"抛回来"的颜色——和天空是蓝色的原因完全相同。
这也是为什么很多婴儿出生时是蓝眼睛，第一年慢慢变深：黑色素还没造出来。
⚠️ The honest version — eye colour is not really one gene
The square above is a teaching model, not the truth. It is the right way to learn
dominant and recessive, and it is wrong about eyes.
Most of the difference between brown and blue does come from one place: a single
letter change near two neighbouring genes, HERC2 and OCA2, on chromosome 15.
But that letter is not the melanin recipe itself — it is a dimmer switch sitting
next to it, turning the melanin gene down. So "blue" is not a broken recipe, it is a
quieter one. Every blue-eyed person alive appears to share that one switch, inherited
from a single ancestor a few thousand years ago.
And it is not alone: at least a dozen other genes nudge the result. That is why the
simple model breaks in three visible ways —
• green and hazel eyes have no box on this square at all;
• brown eyes come in every depth from near-black to light amber, which one B/b square
cannot produce;
• two blue-eyed parents can have a brown-eyed child. It is rare, but it happens,
and the square above says it is impossible.
Real traits are usually like this. One gene, two neat options is the exception —
it is just the exception we teach first, because you cannot understand the complicated
version without it.
Every other cell in your body copies its DNA exactly. Sperm and eggs are the
one exception — and that exception is why you are not a copy of anybody.
Press the button and make a child. You will never get the same one twice.你身体里所有其它细胞都精确复制自己的 DNA，只有精子和卵子例外——
正因为这个例外，你才不是任何人的复制品。按下按钮生一个孩子，永远不会重复。
Why sperm and eggs break the rule — 23 + 23 = 46
You have 46 chromosomes in nearly every cell: 23 pairs. For each pair, one came
from your mother and one from your father. They carry the same genes in the same order — two
editions of the same book.
Sperm and eggs get only 23 — one out of each pair, never both. They are made by a
special kind of division called meiosis, whose whole job is to halve the set.
The reason is arithmetic: if an egg carried 46 and a sperm carried 46, the child would have 92,
the grandchild 184, and life would collapse within a few generations. Halving is what makes
sex possible at all.
At fertilisation one sperm reaches one egg, 23 meets 23, and the count is back to 46 —
but the combination is brand new. That is the moment a genome that has never existed
before, anywhere, starts running.
🔀 Why no two children are alike — two shuffles, not one
Shuffle one: which 23. When a gamete is made, each of the 23 pairs decides
independently which of its two chromosomes goes in. That is 23 coin flips, so a single
person can make 2²³ = 8,388,608 genetically different eggs or sperm. Two parents
together: over 70 trillion possible children — already far more than every human who
has ever lived.
Shuffle two: crossing over. Before the split, the two chromosomes of a pair lie down
side by side, break at matching points, and physically trade the pieces. So the
chromosome you pass on is usually not your mother's or your father's — it is a new
mosaic cut from both. This happens in nearly every pair, every time, at points that are
never quite the same. It is what the two-coloured bars above are showing.
Multiply the two shuffles together and the number stops being meaningful. Every child is,
in the exact sense, unrepeatable — which is also why siblings share about half their
DNA on average, but never the same half.
👴👵 Where your grandparents come in — why their copies reach you
Here is the part people find surprising: the DNA your mother puts into an egg was never
really hers to invent. Her own 23 pairs came the same way yours did — for every pair,
one chromosome came from her mother and one from her father: your grandmother
and your grandfather. The same is true on your father's side.
So when your mother makes an egg, the two shuffles above are shuffling your grandparents'
chromosomes. Independent assortment picks, pair by pair, whether grandma's copy or
grandpa's copy goes in; crossing over then trades pieces between the two. The single
chromosome you receive is usually a mosaic of both grandparents, cut at points that
will never repeat. That is the whole reason a grandparent's DNA comes down through a
parent to you — the parent is a shuffler, not a wall.
It is also why a trait can seem to skip a generation. Your mother might carry your
grandmother's copy without ever showing it — a hidden g from the last lesson — and
hand that exact copy to you, where it finally appears. You can have your grandfather's eyes
even if neither of your parents does.
On average you carry about a quarter of each grandparent's DNA — but "on average" is
doing a lot of work. The shuffle means it might be a third from one grandparent and a tenth
from another, and a few whole chromosomes can pass down untouched. Go back far enough and
some ancestors leave you no DNA at all, even though they are still your ancestors.
令人意外的一点：你母亲放进卵子里的 DNA，其实不是她自己发明的。
她的 23 对也和你一样得来——每一对里，一条来自她母亲、一条来自她父亲，
也就是你的外婆和外公；父亲那边同理。
所以你母亲造卵子时，前面那两次洗牌洗的是你外祖父母的染色体：
独立分配逐对决定进去的是外婆的还是外公的，交换再把两者的片段互换。
你拿到的每一条通常是两位祖辈的拼接。这就是祖辈的 DNA 经由父母传到你身上的原因——
父母是洗牌的人，不是一堵墙。
这也解释了性状为什么会隔代出现：你母亲可能带着外婆的那一份却不表现
（上一课里隐藏的 g），再把这一份原样传给你，在你身上才显出来。
平均来说你从每位祖辈那里各得约四分之一的 DNA，但只是平均——
洗牌可能让某位祖辈多些、某位少些，甚至某些远祖一点 DNA 都没留给你。
👧 Boy or girl — the father's sperm decides
Of the 23 pairs, 22 are the same in everybody. The 23rd pair is the sex pair, and it is
the one pair that can be mismatched: XX or XY.
A mother is XX, so every egg she makes carries an X — she has nothing else to
give. A father is XY, so half his sperm carry X and half carry Y.
Whichever one arrives first sets the result: X + X = girl, X + Y = boy.
The mother cannot influence it, and neither can the father — it is decided by which of
two equally common sperm gets there, which is as close to a coin toss as biology gets.
The Y is tiny and carries very few genes; its main job is a single switch that starts
male development. The X is large and full of ordinary genes that have nothing to do with
sex — which is why some conditions, like red–green colour blindness and haemophilia, show up
far more often in boys: a boy has only one X, so a recessive fault on it has no second copy
to hide behind. It is the G g rule from the last lesson, with the safety net removed.
Identical twins are not two children who happen to match. They are one
fertilisation: a single egg met a single sperm, made one new genome, and then the tiny ball of
cells split in two in the first days. Both halves carry the same DNA, which is why they
are the same sex and look alike. Nothing about the parents causes it — it appears to be
chance, and it happens at about the same rate everywhere in the world.
Fraternal twins are the ordinary process run twice at once: the mother released
two eggs, and two different sperm fertilised them. Genetically they are simply
siblings who share a birthday — about half their DNA, and they can be different sexes.
Even identical twins are not perfectly identical for long. Copying errors accumulate
separately in each of them from the first division onwards, and — as the previous lesson
showed — they open different pages at different times. Their fingerprints differ,
because a fingerprint is shaped by how a hand happened to press against the womb.
Almost everything above is a fifty-fifty deal. One thing is not.
Your cells contain tiny power plants called mitochondria, and they carry their own
small loop of DNA — 37 genes, entirely separate from the 46 chromosomes. The egg is a
large cell and brings thousands of mitochondria into the new embryo. The sperm is little more
than a nucleus with a tail, and the few it carries are destroyed on arrival.
So mitochondrial DNA comes from your mother alone — and from her mother, and hers,
unbroken, all the way back. It is never shuffled and never crossed over, so it changes only
by slow mutation. That makes it a clock: it is how researchers trace maternal lines across
tens of thousands of years, and it is why the phrase "mitochondrial Eve" exists.
Fathers pass on none of it.
The square in the previous lesson can now be read properly. It looked like a table of children.
It is really a table of gametes.
The letters along the top are the two alleles the father can put into a sperm. The
letters down the side are the two the mother can put into an egg. Each box is one
possible meeting — one particular sperm reaching one particular egg.
So "3 out of 4" was never a promise about four children. It is the chance at each
conception, and chance has no memory: three brown-eyed children in a row does not make
the fourth more likely to be blue, exactly as three heads in a row does not bend the next
coin. A family of four can easily come out 4–0. The ratio only becomes visible across
thousands of families — which is precisely how Mendel found it, counting some
28,000 pea plants rather than looking at one.
You cannot give an order you cannot say. Learn these first.
先学会这些词，才能下指令。
🎤 Speaking drill —
Press the mic, then say all five words out loud.
按下麦克风，把五个词都说出来。
Your words will appear here…
🌍 The Crisis Board
Seven problems are killing Earth. You cannot fix them all.
七个问题正在毁灭地球。你们不可能全部解决。
🧑🔬 The Volunteers
Eight humans said yes. Read their data — the data tells you what their DNA can do,
and which problem each one is best for.
八个人自愿参加。读他们的数据，看看每个人最适合哪个问题。
✍️ Commit your choice
Doctor 1 picks a human — the problem it is best for is matched automatically.
Then Doctor 2 picks a different human.
医生1选一个人，系统自动匹配最适合的问题；医生2再选另一个人。
🙈
🎧 Listen to the command. Then say it back.
先听指令，再复述出来。
Everything you say appears here…
🛸 Final transmission
Last task — argue about it:
“Whose modification was better? I think ______, because ______.”
“If we play again, I will insert the ______ gene instead.”