It might have been a weird choice for a 20-year-old Welsh boy with accent anxiety to then do a masters in Welsh sheep genomics. I promise you it wasn’t my fault. Although I’ve been ruminating on that lately, especially as I think about accent anxiety and what it means for me to be Welsh in London. I’ve come to realise that while I was doing menial work pushing files through pipelines as a masters student, there were lessons I was exposed to in why this gruntwork becomes what science is.

My involvement starts after a terrible lab placement in 2018 where nothing worked and my protocols kept going wrong. Explode in my face kind of wrong. I thought long and hard about what would make me employable and decided from there the world of bioinformatics held the most promise.

Most reading this will understand for early career scientists, the projects you take are mostly based on skills which take time to hone, rather than the domain knowledge which can be easily mopped up. But even so, you could say I was mis-sold on the project. When I interviewed for it, I was the subject was Russian cattle breeds (as you can guess, scientific collaborations between UK and Russia have halved from 2,366 in 2021 to 1084 in 2025, that’s out of 200K papers in 2025 in general). the supervisor was a Russian national. Then my surprise when I read South_Welsh_Mountain.FASTA on the screen. On account of that accent anxiety, he must have mentioned he was previously based at Aberystwyth. No hard feelings though, I had a great time, and agree that domestication, alongside language, is one of the most important events in human history.

My point being that what I was doing was researching valuable mutations that have been selected for in local breeds. Usually breeds of animal in commercial senses are bred for productivity, and with that comes susceptibility to disease and stress that their “wilder cousins” have adapted to. Think of the Texel, which is now crossed with other breeds to increase hardiness. So what was important was adaptation to local environment, but also how sheep play their own role in the local environment.

We split into lowland breeds: productive but not hardy, and upland: hardy but not productive. Any stark differences between the two that have been selected for, and aside from genetic bottlenecks you’d expect in inbred animals. I found some interesting things: signals for genes in mitochondrial biogenesis, oxidative stress pathways. Due to 2020 happening, I was able to verify none of this in the lab to prove if these were activating or suppressing. Arguably these are the missing pieces of the puzzle, but in that we build up databases of potentially useful mutations with seemingly beneficial effects.

The reason we do this (or at least what is put on the grant applications) is to understand these mutations, transplant them to other breeds with the view of making them much hardier to the oncoming environmental stress iun a warming world. The way it’s done is by researching the literature around the mutations your bioinformatics pipeline suggest are under selection and then identify why it might be important, and see what fits the story. It’s not the most rigorous approach, but people are doing this all around the world: Russian sheep breeds, Chinese sheep and Brazilian sheep to name some examples that have cited my work.

Going back to the quote suggesting there are two types of science: Physics and stamp collecting, I certainly don’t agree with the quote, but I understand the sentiment. Even so, the things I’ve seen in my working life, I don’t agree with the sentiment. The merit of stamp collecting are resulted in RNA vaccines ready when there’s a pandemic, seed banks are full in famine, and LLMs can write your Christmas Cards.

Modern science is not the story of leaps, it is the story of incremental steps that lead to a maturation and a strengthening community researchers. It’s about a community identifying the gaps in knowledge and filling them, and really that’s what the scientific method’s all about. Knowing the gaps can help you decide where you go next.

And then comes my feelings on AI scientists. I’m hearing a lot about autonomous labs, AI models that will be able to observe phenomena, test hypotheses and change accordingly. The capabilities of these AI scientists is largely unknown, to which, my question is: will they be doing the science or the stamp-collecting? We can hope that they could do the science. The US is spending $5b to find out.

Based on my understanding of LLMs my feeling is more skeptical and the best place for AI for science is in inference. Their whole USP is that they can make inferences that people can’t, but they can only do that from their own reference point in their training data. If that reference does not exist. Will it look at results and reply “that’s funny,”?

Furthermore, this gruntwork is what the models get trained on. Repositories like what I mention with my sheep example make things like protein language models possible. Which was the basis of my PhD. It’s the reason why big pharma companies are raiding stockpiles of lab books for proprietary data to train their own models: again, I think this could all be federated to achieve more generalisable models quicker, without surrendering training data, but companies be companies.

What I do think is possible, is more akin to the Modern Synthesis: introducing the mechanics of Gregor Mendel’s inheritance laws with Darwinian evolution. A romanticised story of a chance journal reading on a train. Now the scientific corpus is more searchable, perhaps less obvious inferences can be made. But these are still inferences, rather than discoveries.

To be clear though, this isn’t a bad thing. The whole point of this piece is to defend stamp-collecting and champion its importance in filling gaps in our knowledge. Autonomous systems doing this kind of work, frees up people to do more science. AI can suggest how to speed up workflows and identify the most relevant literature to base future hypotheses on.

Back to sheep, as the world heats and challenges of the future become challenges of today, we will need every chance of overcoming that science allows. I certainly welcome the idea of automating science, but only if it brings benefits of allowing the people to be the ones who make the breakthroughs, and that they remain robust and reproducible. While the breakthroughs are the things that attract the glam-mags, Stamp-collecting remains as important, if not more important than ever. After all, these models are trained on it.