Aubin Ramon

speaker

346 appearances 1 recordings 1 series first heard Jul 2026 last heard 6 Jul

Aubin Ramon’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 1 in all, peaking in Jul 2026 with 1.

Appearances

newest first · ▶ plays the moment

Talking Techniques · AI & Antibodies miniseries | Nanobody thermostability prediction and the great data challenge · 6 Jul 2026 podcast

And because we have small data, then we're gonna have

A lack of diversity in into the representation of the data.

And a lack of also standardized and high quality data points.

Because only like a few people from different labs, from different research institutions are gonna measure their protein in their own buff environment with their own um properties and uh condition, experimental setup conditions.

Uh and then when you try to train a model among this limited data set, so small amount.

um small quality, uh little quality and uh poor diversity is gonna struggle.

So how we did to

overcome this is um we um started by characterizing more nanobodies because when I started this project, so at the beginning of my PhD four five years ago, we only had less than a hundred data points of nanobody femor stability, which is

spoiler alert nothing for a machine learning model.

And so what we did is that okay, so let's characterize more.

So we took um the PDB of nanobodies, we selected a hundred diverse sequences, we

uh order the expression and purification and then we measured the thermostability in the same buffer, but gave rise to a higher quality data set, around two hundred, two hundred fifty data points, try to train a model wasn't enough.

But luckily enough, at the same time, we had um another lab that published a data set named NBFEMO that tried to pass throughout the literature.

So in that paper is VS Fermo stability for V SND.

In that other paper, there are a few for a few variants.

It collected everything and rise to around 400, 500 new data points.

So altogether I had 600 data points.

Which doesn't sound which might sound a lot for some people, but actually it's nothing for machine learning.

Um and

Showing 21–40 of 346 · page 2 of 18