System.Xml.XmlElement
A few years ago I stumbled upon Leash Bio’s Substack , and was hooked.
几年前，我偶然发现了 Leash Bio 的 Substack，并被迷住了。
Be it their Kaggle competition that showed that no, we really haven’t solved protein-ligand binding , or their crazy discovery that knowing a molecule’s author is perhaps all you need to know to predict how good it is .
无论是他们的 Kaggle 竞赛表明，不，我们确实还没有解决蛋白质-配体结合的问题，还是他们的疯狂发现，即了解分子的作者也许就是你预测它有多好所需要知道的一切。
Their substack — or at least Owl Posting ’s 1 deep dive on them — should be required reading for any scientist trying to do chemistry with AI.
他们的子堆栈——或者至少是 Owl Posting 对它们的第 1 次深入探讨——应该是任何试图与人工智能进行化学反应的科学家必读的内容。
So I was super excited to finally be able to sit down with Ian Quigley from Leash Bio, and pick his brain.
所以我非常兴奋终于能够与 Leash Bio 的 Ian Quigley 坐下来，并听取他的意见。
Leash specializes in DNA Encoded Libraries (aka “DELs”), a way of collecting ridiculously high-throughput interaction data between proteins and small molecules.
Leash 专门研究 DNA 编码文库（又名“DEL”），这是一种收集蛋白质和小分子之间高通量相互作用数据的方法。
2 Before I discovered Leash, I had heard nothing but horror stories from DEL users.
2 在我发现 Leash 之前，我只听到 DEL 用户的恐怖故事。
And yet, Ian and friends somehow found a way to tame the beast, extracting clean and reproducible data that is just perfect for large scale modeling.
然而，伊恩和朋友们以某种方式找到了一种驯服野兽的方法，提取干净且可重复的数据，这些数据非常适合大规模建模。
Without teasing too much, the answer comes down to epistemic rigor, and an incredible conviction that you just have to Do The Thing.
无需过多调侃，答案可以归结为认知上的严谨性，以及一种令人难以置信的信念，即你只需要做该做的事。
It gets even better!
它会变得更好！
Near the end of this episode, Ian teases something his team was just playing with when we recorded: an AI playground for agents to do verifiable scientific reasoning.
在本集快结束时，伊恩调侃了我们录制时他的团队正在玩的东西：一个供代理进行可验证的科学推理的人工智能游乐场。
They discovered the data they had spent the last few years collecting could provide tens of thousands of “lab-in-the-loop” simulations, just what an agent needs for scientific RL.
他们发现过去几年收集的数据可以提供数以万计的“实验室在环”模拟，这正是智能体进行科学强化学习所需的。
Two months later, it’s matured into Minos, and today they’re releasing a full preprint , blog , and open dataset .
两个月后，它已经成熟为 Minos，今天他们发布了完整的预印本、博客和开放数据集。
We’ll be releasing a lightning podcast in a few days with an update.
几天后我们将发布带有更新的闪电播客。
Stay tuned!
敬请关注！
We had a blast with Ian, probably one of our most fun episodes to date.
我们和伊恩玩得很开心，这可能是我们迄今为止最有趣的剧集之一。
For both complete chemistry novices and expert AI x Bio practitioners, we think you’ll get something out of this.
对于化学新手和人工智能 x 生物专家从业者来说，我们认为您会从中有所收获。
Why DELs traumatize everyone who touches them, and how Leash puts dozens of barcodes on each molecule to catch proteins that bind the DNA instead of the drug.
为什么 DEL 会伤害每个接触它们的人，以及 Leash 如何在每个分子上放置数十个条形码以捕获与 DNA 而不是药物结合的蛋白质。
Ian’s rule that biology only scales two ways: taking pictures and sequencing DNA.
伊恩的规则是，生物学只能通过两种方式进行扩展：拍照和 DNA 测序。
The Kaggle competition : about 4B measurements from a basement and about 2,000 teams, with nobody beating chance on an unseen chemical library.
Kaggle 竞赛：来自地下室和大约 2,000 个团队的大约 4B 测量，没有人在看不见的化学库上击败机会。
Also why Leash quietly feared co-folding would make them obsolete.
这也是 Leash 悄悄担心共同折叠会让它们过时的原因。
Clever Hans in ChEMBL : models can guess a molecule’s chemist from its structure, and “who made it” predicts activity about as well as the structure does.
ChEMBL 中的聪明汉斯：模型可以根据分子的结构猜测其化学家，而“谁制造了它”可以像结构一样预测活性。
Hermes , trained only on Leash’s own data, keeping pace with Boltz on public benchmarks.
Hermes 只根据 Leash 自己的数据进行训练，在公共基准上与 Boltz 保持同步。
Plus the BRD4 molecule Boltz had already seen in the PDB.
再加上 Boltz 已经在 PDB 中看到的 BRD4 分子。
Breaking the pocket : engineering protein mutants at scale after catching Hermes memorizing “nicknames” for proteins instead of binding sites.
打破口袋：在发现 Hermes 记住蛋白质“昵称”而不是结合位点后，大规模设计蛋白质突变体。
The agent sandbox : LLMs that reason about SAR in plain English, hunting for the best binders among 6.5M molecules on a budget of 100.
代理沙箱：法学硕士用简单的英语推理 SAR，以 100 的预算在 650 万个分子中寻找最佳结合剂。
Going from a DEL hit to good mouse PKPD in six months on an inflammation program, and the case for pills over antibodies on cytokine targets.
在炎症项目中，六个月内从 DEL 命中到良好的小鼠 PKPD，以及针对细胞因子靶标的药物优于抗体的情况。
What stops a better-funded lab from outscaling them, and Ian’s by-fiat wish for massively parallel purification.
是什么阻止了资金更充足的实验室扩大规模，以及伊恩对大规模并行纯化的必然愿望。
As always, the secret here is to map your problem to a combinatorial library and a readout using Next Generation Sequencing.
与往常一样，这里的秘密是将您的问题映射到组合库并使用下一代测序进行读数。
We talk about this in depth in the podcast!
我们在播客中深入讨论了这一点！