David Younger

Check out our new AlphaBind preprint demonstrating our industry-leading antibody optimization model! Also excited to share that we have measured our BILLIONTH protein-protein interaction, providing the crucial training data to predict protein binding. biorxiv.org/content/10.110…

@alpha29er @BiologyAIDaily @AAlphaBio Check again! [github.com/A-Alpha-Bio/al…] bioRxiv approval was faster than we expected!

AlphaBind: a Domain-Specific Model to Predict and Optimize Antibody-Antigen Binding Affinity @AAlphaBio 1. AlphaBind is introduced as a transformer-based model trained on 7.5 million antibody-antigen binding data points, specifically designed for affinity prediction and optimization in therapeutic antibody development. It allows efficient generation of high-affinity antibody candidates with only a single round of optimization. 2. The model leverages ESM-2nv embeddings and transfer learning, resulting in impressive affinity gains across various antibodies. It achieved affinity improvements up to 74-fold for antibodies like AAB-PP489 and demonstrated high target specificity, with 92-97% of candidates binding their intended targets effectively. 3. AlphaBind’s fine-tuning pipeline includes affinity-guided developability engineering, where it optimizes binding while addressing biophysical properties, like solubility and immunogenicity. This resulted in optimized anti-TIGIT variants with high binding affinity and improved developability. 4. Ablation studies confirmed the critical roles of ESM-2nv embeddings and pre-training data, showing that removing these components led to a significant drop in model performance. 5. AlphaBind showcases the power of large-scale data and AI-driven antibody optimization, with potential to expedite therapeutic antibody design and reduce reliance on iterative experimental screening. @randolph380 @DavidAYounger @lahman_miranda @Kerry__McGowan 💻Code: github.com/A-Alpha-Bio/al… 📜Paper: biorxiv.org/content/10.110… #AntibodyEngineering #Bioinformatics #MachineLearning #AI #Biopharma

The glue degraders go.nature.com/3IqK9pT

Excited to share our new paper using AlphaSeq data to enable predictive ML models for antibody engineering with our wonderful collaborators at @Google and @LumenBio biorxiv.org/content/10.110…

It was a pleasure speaking with @CPicardo from @MadronaVentures about @AAlphaBio and building a company at the intersection of life science and data science. Thank you for featuring A-Alpha Bio in the first Founded and Funded podcast of 2023!

We are partnering with Bristol Myers Squibb to discover targets for molecular glues! Very excited to announce this collaboration and even more excited to show what we can do together.

A-Alpha Bio uses yeast and machine learning to help find new therapeutics. Their ultimate aim is to find the right therapeutic protein through in silico prediction. And they harness the yeast mating system to get the critters to fuse!

Congrats to @AAlphaBio on their Series A. We are excited to be working with you! Curious what an Intersection of Innovation company does? - read more here from @mattmcilwain @CPicardo bit.ly/3tnLo0J

@xontogeny @AAlphaBio Very excited to be working with Ben and the Xontogeny team - and very happy to be the first investment from the new PXV Fund II!

“We’re thrilled to support A-Alpha Bio’s next phase of growth as they continue to set new standards for protein and molecular glue research” - @Xontogeny’s Ben Askew, on selecting @AAlphaBio for the first investment from PXV Fund II. prweb.com/releases/a_alp…

I am very excited to share that @AAlphaBio has raised a $20M Series A, led by @MadronaVentures and with new investors @xontogeny and @Lux_Capital! We're on a very exciting trajectory to massively impact drug discovery with synthetic biology and machine learning!

Very excited to share our first @AAlphaBio preprint, where we map how leading therapeutic COVID-19 antibodies are affected by thousands of CoV-2 variants using #alphaseq. Which antibody is best against emerging SARS-CoV-2 variants? Read the preprint to find out!

Which therapeutic antibodies retain binding against emerging CoV-2 variants? Our latest pre-print is out! 33 monoclonal antibodies (including 9 in clinical trials) tested for binding against 1000's of CoV-2 variants. biorxiv.org/content/10.110… [thread 1/4]

Very excited to be participating!

We are thrilled that the @NSF is continuing to support our development of AlphaSeq for molecular glue discovery with a Phase II SBIR grant! Read more about it below:

We are very excited to announce that we have been awarded a Phase II SBIR from the NSF to continue developing AlphaSeq for molecular glue discovery!! aalphabio.com/post/a-alpha-b… Special thanks to @b0yle of GeekWire for writing a great article that gives more detail. #AlphaSeq

Our immune systems do a good job of making antibodies, but these don't always make for great drugs. We're using GANs to create humanoid antibodies, designed the way our bodies would make them, but optimized for therapeutic purposes. A new platform for faster, better drugs.

A-Alpha Bio is hiring! We are looking for a talented PhD structural biologist with experience in targeted protein degradation to join A-Alpha Bio as a Scientist. See below for more info and to apply! linkedin.com/jobs/cap/view/…

Very excited that microSPLiT, our approach to high-throughput scRNA-seq of bacteria is now out. Led by the incredible @annabiosys and @LeandraBrettner with @RocoCharlie @dna_rosenberg Will DePaolo, Luana Paleologu, Matt Hirano, Ryan Kibler, A. Carignano biorxiv.org/content/biorxi…

We are looking for an experienced Research Associate to join our R&D team in #Seattle! This role involves working with AlphaSeq and cutting-edge tools from synthetic biology and NGS for drug discovery applications. Apply or refer a friend: bit.ly/33lFAHl