You Inherit the Letters — and the Notes Scribbled in the Margins
June 21, 2026 · Erwin Schrödinger, What Is Life?~5 min read
In school you learned a clean story: heredity is a string of A, T, C, G — four letters, copied from parent to child, and that string is you. It is a beautiful story, and it is mostly true. But a new study of mice has caught inheritance smuggling something extra across the generations: not new letters, but notes pencilled in the margins of the page — annotations that can pass down, and that sometimes appear in the child's copy from nowhere at all.
A book whose margins are alive
Picture your genome as a very long book. The text is the DNA sequence — the letters everyone means when they say "your genes." But a book has more than text. Some words get a highlighter, some lines a note in the margin: read this loudly, skip that, this one matters. In cells those marks are real chemistry. The most studied is methylation — small chemical tags that clip onto the DNA and quietly switch genes on or off without touching a single letter. The text stays identical; the annotations change how it is read. Geneticists call this the epigenetic layer. I think of it as the margins of the code.
What the mice actually showed
Researchers at Johns Hopkins, publishing in Nature Genetics in May 2026, followed those margin notes through three generations of mice — grandparents, parents, pups — reading the methylation pattern at each step. Most of it behaved exactly as a century of genetics predicts: marks sorting like Mendel's peas, neat and lawful. But not all. In 522 cases the inheritance broke Mendel's rules — about 7% of the methylation patterns refused to follow the clean script. Two findings stand out. One is a paramutation: one version of a gene heritably altering its partner, a ghostly influence seen before in maize and in flies but, until now, not nailed down in a mammal. The other is stranger. The team found "emergent" methylation — margin notes that turned up in the pups while being present in neither parent. An annotation in the child's copy of a book that was in neither parent's copy.
A Johns Hopkins study (Nature Genetics, May 2026) tracked DNA methylation — chemical tags that switch genes on or off without changing the DNA letters — across three generations of mice, and found 522 cases (about 7%) inherited in non-Mendelian ways, including a mammalian paramutation and "emergent" marks present in neither parent. Read: the code-script has margins, and the margins are heritable too. Framework: Erwin Schrödinger, What Is Life? Popular-science interpretation; mice, a minority of marks, not a rewrite of genetics.
Schrödinger guessed the text. He couldn't see the margins.
This sits strangely close to a book written before anyone knew DNA existed. In 1944 the physicist Erwin Schrödinger asked, in What Is Life?, a question biology hadn't framed so sharply: how can heredity be stored stably enough to last across generations, yet be intricate enough to specify a whole organism? His answer was a leap. The carrier had to be a molecule he called an "aperiodic crystal" — orderly enough to endure, irregular enough to hold a vast message. A stable molecular code-script. A decade later Watson and Crick found his crystal: the double helix, and the letters running down it were the code-script made real. Schrödinger got the text right before the text was found. What he could not have guessed is that the script would come with a second layer — chemical annotations sitting above the letters, riding along into the next generation, and on rare occasions written fresh in a child the parents never carried. The code-script, it turns out, has margins.
Schrödinger predicted a stable molecular code-script, and DNA proved him right about the letters. The surprise is that the script has margins — and the margins are heritable too.
You inherit the text. Sometimes you also inherit the notes.
How to read this without overclaiming
Here is where a naturalist has to hold two things at once: wonder, and rigor. The wonder is real — inheritance is layered in a way the textbook diagram never showed. But so are the limits, and they matter. This is mice, not people. It is roughly 7% of marks, not the whole genome — the other 93% behaved themselves. And epigenetic inheritance has a long, bruising history of dramatic claims that later dissolved into confounds: a mark that looked inherited turned out to be the mother's diet, or a contaminant in the cage. So no, this is not "Lamarck was right" and your gym habit will not reach your grandchildren. It does not rewrite genetics; the central story — letters, copied, sorted by Mendel — still carries the day. What it does is complicate a picture that was a little too clean. A measurable minority of heredity travels in the margins, and at least in mice, the margins can carry surprises.
The next time a headline shouts that scientists have "rewritten the laws of inheritance," reach for the margins-of-a-book image and ask three small questions. Is this in mice, or in us? Is it most of the genome, or a careful 7%? And has the obvious confound — diet, environment, a contaminated cage — actually been ruled out? Schrödinger's instinct was that life hides its inheritance in a stable script, and he was right. The newer, humbler lesson is that the script was never only its letters. We are still learning to read in the margins.
Framework paraphrased from Erwin Schrödinger, What Is Life? (1944), and its idea of heredity as a stable molecular "code-script" (the "aperiodic crystal" that anticipated DNA). Real basis: a Johns Hopkins mouse study in Nature Genetics (May 2026), funded by NIH/NSF, which tracked DNA methylation across three generations and reported 522 cases (~7% of methylation patterns) inherited in non-Mendelian ways, including a mammalian paramutation and "emergent" methylation present in neither parent. Popular-science reflection, not a research claim; findings are in mice, concern a minority of marks, and epigenetic-inheritance results have a history of confounds — this complicates, but does not rewrite, classical genetics.
把你的基因组想成一本极长的书。正文是 DNA 序列——人们说"你的基因"时,指的就是这些字母。可一本书不只有正文。有些词被荧光笔划过,有些行旁边写着批注:这段大声读,那段略过,这一句很重要。在细胞里,这些标记是实打实的化学。研究得最多的是甲基化——一些小小的化学标签夹在 DNA 上,一个字母都不动,却悄悄把基因开关切掉或打开。正文一字未变,批注却改变了它被怎样阅读。遗传学家把这一层叫表观遗传层。我更愿意叫它:密码本的页边。
约翰斯·霍普金斯大学的研究(《自然·遗传学》,2026 年 5 月)追踪了 DNA 甲基化——一种不改变 DNA 字母、却能开关基因的化学标签——横跨三代小鼠,发现 522 例(约 7%)以非孟德尔的方式遗传,其中包括哺乳动物身上的副突变,以及父母双方都没有、却在后代涌现的标记。解读:密码本有页边,而页边也会遗传。框架:薛定谔《生命是什么》。本文为科普解读;限于小鼠、限于少数标记,并非改写遗传学。
薛定谔猜中了正文,却看不见页边
这件事,奇异地贴近一本在人类知道 DNA 之前就写成的书。1944 年,物理学家薛定谔在《生命是什么》里,提出了一个生物学当时还没问得这么锋利的问题:遗传如何既稳定到足以跨越世代留存,又精巧到足以指定一整个生物体?他的回答是一次飞跃。那个载体必须是他称作"非周期晶体"的分子——规整到足以存续,又不规则到足以承载一条庞大的讯息。一份稳定的分子密码本。十年后,沃森与克里克找到了他的晶体:双螺旋,而沿着它排列的字母,正是那份密码本成了真。在正文被找到之前,薛定谔就猜对了正文。他想不到的是,这份密码本还附带第二层——坐在字母之上的化学批注,搭着便车进入下一代,偶尔还会在父母从未携带的孩子身上被新写出来。原来,这密码本是有页边的。
薛定谔预言了一份稳定的分子密码本,而 DNA 在"字母"这件事上证明他是对的。意外在于:这份密码本有页边——而页边也会遗传。
あなたのゲノムを、とても長い一冊の本だと思ってほしい。本文は DNA 配列——人が「あなたの遺伝子」と言うときに指す、あの文字たちだ。だが本には本文以外のものもある。ある語には蛍光ペン、ある行には余白の覚え書き。ここは声に出して読め、そこは飛ばせ、この一文が肝心だ、と。細胞の中では、その印は正真正銘の化学だ。最もよく調べられているのがメチル化——小さな化学標識が DNA に留め金のように付き、一文字も変えずに遺伝子をそっとオン・オフする。本文は同一のまま、書き込みだけが読まれ方を変える。遺伝学者はこの層をエピジェネティック層と呼ぶ。私はこう呼びたい——暗号文の余白、と。
これは、人類が DNA の存在を知る前に書かれた一冊に、奇妙なほど近い。1944 年、物理学者シュレーディンガーは『生命とは何か』で、当時の生物学がまだこれほど鋭く立てていなかった問いを投げた。遺伝はどうして、世代を越えて残るほど安定でありながら、生物まるごとを指定できるほど精緻でいられるのか。彼の答えは飛躍だった。その担い手は、彼が「非周期結晶」と呼ぶ分子でなければならない——存続できるほど整い、膨大な伝言を抱えられるほど不規則な分子。安定した分子の暗号文だ。十年後、ワトソンとクリックが彼の結晶を見つけた。二重らせん、そしてそれに沿って並ぶ文字こそ、現実となった暗号文だった。本文が見つかる前に、シュレーディンガーは本文を言い当てた。彼に予想できなかったのは、その暗号文に第二の層が付いてくることだ——文字の上に座る化学的な書き込みが、次の世代へ便乗し、ときには親が一度も携えなかった子に新たに書き込まれる。暗号文には、余白があったのだ。