Daniel Wigh

Co-Founder @ ReactWise (YC S24)
PhD from University of Cambridge

[Daniel Wigh @DanielSWigh]

@cremieuxrecueil Whoops haha

[Daniel Wigh @DanielSWigh]

I made the @Forbes 30 Under 30 list 🤯

[Y Combinator @ycombinator]

Congrats to the team @ReactWise on the $3.4M round!

ReactWise's copilot for chemical process optimization enables AI-driven robotic laboratories and speeds up pharmaceutical manufacturing R&D.

https://t.co/Gu6WU1JI26

[Arda Göreci @ArdaGoreci]

🚀Excited to announce: Open-source AlphaFold3 implementation! 🚀

I am thrilled to announce one of the models we have been building for the last 8-weeks at Ligo - an open-source implementation of DeepMind’s frontier model, AlphaFold3! Here’s what we have learned, a thread (1/11):

[ReactWise @ReactWise]

We are excited to share a video with insights into our ambitions at ReactWise to drive efficiency and innovation in pharma & biotech. Reach out to us via [email protected]

https://t.co/WyY0WEOeF7

Video credit: Nathan Pitt ©University of Cambridge

[Daniel Wigh @DanielSWigh]

@ycombinator @ReactWise

[Daniel Wigh @DanielSWigh]

ReactWise just launched on @ycombinator's Launch YC!

ReactWise - AI co-pilot for wet lab chemistry. Check them out: https://t.co/g3MSrQkafI

[Daniel Wigh @DanielSWigh]

@marcos_felt @alexanderpomb @LapkinLab @SynTech_CDT @ReactWise @iDMTcambridge @Cambridge_Uni

[Daniel Wigh @DanielSWigh]

Cleaning chemical reaction data sets has never been easier! Check out ORDerly, our new paper in JCIM, about our open source pipeline for cleaning chemical reaction data sets stored in Open Reaction Database (ORD) https://t.co/IsvOEH1VVk

[Daniel Wigh @DanielSWigh]

We also created three new benchmarks: ORDerly-forward (forward prediction), ORDerly-retro (retrosynthesis), and ORDerly-condition (the first condition prediction benchmark).

[Daniel Wigh @DanielSWigh]

You can find our algorithm explained in detail in our paper: https://t.co/vXlYjqckFt #chemistry #protodeboronation #computationalchemistry #boronicacid #crosscoupling

[Daniel Wigh @DanielSWigh]

Do you work with boronic acids? If so, check out this web app we built based on a recent paper of ours! https://t.co/ziCgrZ4PFW
Simply input the SMILES string of a boronic acid, and it will show you the predicted rate of protodeboronation (boronic acid degradation)!

[Daniel Wigh @DanielSWigh]

The rate prediction is made as a function of pH in 50/50 water/dioxane. If our dataset does not contain your queried molecule, it will find the most similar one to show instead (by tanimoto similarity).

[Daniel Wigh @DanielSWigh]

Special thanks to my supervisors Prof. Lapkin and Prof. Goodman, and our collaborators at UCB Pharma.

[Daniel Wigh @DanielSWigh]

Excited to share our work on predicting the rate of the protodeboronation reaction! To anyone working with boronic acids, you can find predicted rates of protodeboronation for 50 academically and industrially relevant molecules in the SI of our paper!

https://t.co/YtQZOTxBvF

[Daniel Wigh @DanielSWigh]

In this work we predict protodeboronation rates by considering each mechanistic pathway separately, and calculating characteristic energy differences using Density Functional Theory (DFT), without the need for further experimentation.

[Daniel Wigh @DanielSWigh]

Protodeboronation is a significant degradation pathway for boronic acids, though understanding how quickly it occurs (and how to avoid it) remains a challenge, and relies on expert knowledge and wet-lab experiments.

[Lapkin Lab @LapkinLab]

Our new paper creates a workflow for predicting protodeboronation rates from a SMILES string! Amazing work from @DanielSWigh! #chemtwitter
https://t.co/zaB2QRN4bj

[Daniel Wigh @DanielSWigh]

Thank you to my supervisors @AlexeiLapkin and @goodman_j for all your excellent support and feedback, and thank you @SynTech_CDT and @ucb_news for funding this work!

[Daniel Wigh @DanielSWigh]

My paper ‘A review of molecular representation in the age of machine learning’, has just been published! it explores how to describe molecules in a way that is understandable to computers and compatible with machine learning workflows. Check it out here: https://t.co/D3mpq3kNm3

[Daniel Wigh @DanielSWigh]

Might be interesting if you want to learn more about string (e.g. InChI/SMILES), feature based (e.g. fingerprints) and computer learned (e.g. VAEs) representations.