NVIDIA brings Alchemab into the Cambridge-1 supercomputer project


Cambridge biotech pioneer Alchemab Therapeutics has been chosen by US tech giant NVIDIA to harness the power of  the UK’s most powerful supercomputer, Cambridge-1. 

Alchemab will use the NVIDIA DGX SuperPOD supercomputing cluster, powered by NVIDIA DGX A100 systems, to gain greater understanding and insights from its extensive neurology and oncology datasets. 

Based at Babraham Research Campus, Alchemab focuses on the discovery and development of naturally-occurring protective antibodies and immune repertoire-based patient stratification tools. The Cambridge-1 supercomputer is housed at Kao Data Centre in Harlow.

Alchemab CEO Douglas Treco said: “We are honoured to collaborate with NVIDIA to advance our work applying machine learning to the prediction of antibody structure and function. 

“Using Cambridge-1, Alchemab will vastly accelerate our capabilities and we are excited about the potential to collaborate with NVIDIA’s world-leading team to better understand the language of antibodies.”

Craig Rhodes, EMEA Industry Lead for Healthcare and Life Sciences at NVIDIA, added: “Cambridge-1 enables the application of machine learning to help solve the most pressing clinical challenges, advance health research through digital biology, and unlock a deeper understanding of diseases.

“The system drives workloads that are scaled and optimised for supercomputing and will help extraordinary organisations like Alchemab, a member of the NVIDIA Inception program, to further their research on antibodies and other protective therapeutics for hard to treat diseases.” 

An individual’s antibody repertoire encodes information about past immune responses and potential for future disease protection. Alchemab believes that deciphering information stored in these antibody sequence datasets will transform the fundamental understanding of disease and enable discovery of novel diagnostics and antibody therapeutics.

Using self-supervised machine learning, Alchemab has developed antibody-specific language model AntiBERTa (Antibody-specific Bi-directional Encoder Representation from Transformers), a 12-layer transformer model which provides a contextualised numeric representation of antibody sequences. AntiBERTa learns biologically relevant information and is primed for multiple downstream tasks which are improving our understanding of the language of antibodies.

Alchemab has developed a highly differentiated platform which enables the identification of novel drug targets, therapeutics and patient stratification tools by analysis of patient antibody repertoires. 

The platform uses well-defined patient samples, deep B cell sequencing and computational analysis to identify convergent protective antibody responses among individuals that are susceptible but resilient to specific diseases.

Alchemab is building a broad pipeline of protective therapeutics for hard-to-treat diseases, with an initial focus on neurodegenerative conditions and oncology.

The highly specialised patient samples that power Alchemab’s platform are made available through valued partnerships and collaborations with patient representative groups, biobanks, industry partners and academic institutions.

It’s almost a year since Alchemab completed an $82 million Series A financing round – the largest European Series A round last year and in the top 10 for the largest Series A rounds in Europe. It was backed by an international investment syndicate.

Article sourced from