Drug Discovery

Get in touch



Accelerated Drug Discovery

A novel hybrid platform combining Quantum inspired technology, Machine Learning and advanced physics simulations improving both speed and chemical diversity in designing novel virtual candidates. Our platform creates a target bespoke virtual library, containing billions of molecules, all designed to hit the Clinical Research Organisation’s specific target, achieving de novo molecule discovery while minimising timescales.


The Technology Powering the Small Molecule Drug Discovery Platform

At the heart of Fujitsu’s Drug Discovery Offering is the Digital Annealer. The Quantum-inspired Digital Annealer solves combinatorial optimisation problems orders of magnitude faster than any other currently available alternatives. It is based on custom hardware, is production ready and incorporates Quantum-like processes and algorithms. Forward looking customers are using the Digital Annealer (DA) to help prepare their computational strategy for the Quantum Revolution.

Supporting our clients

We build a new, target bespoke, virtual combinatorial library (billions+) at the initiation of a project. This can then be sampled by the DA, optimising the molecules toward our drug design relevant criteria, including proprietary QUBO (Quadratic unconstrained binary optimisation) based objective functions - Synthetic Accessibility and Electrostatic Similarity.

Figure 1: Platform overview



Figure 2: Post-DA cascade process overview

The DA-based enumeration of a targeted library of billions of molecules yields a refined list of molecules optimised toward both synthetic accessibility and electrostatic similarity. This output is further filtered and ranked in a client bespoke manner, utilising industry-leading Machine Learning algorithms for ADMEt (absorption, distribution, metabolism, excretion, and toxicity) prediction, followed by tailored Ligand & Structure based virtual screening. From here, if applicable to the given target, the molecules are subject to advanced physics simulations (e.g. FEP or QM/MM), generating a more robust prediction of binding affinity. The output is highly configurable to client need and capability: from a short-list of quality molecules prioritised for synthesis and in vitro testing, to a fully managed wet lab procedure.

main.css
life_science_casestudy
media.css
above-rack-image
book-a-tour-image
customisation-image
eight-mile-image
fast-seamless-image
homebush-image
malaga-image
melbourne-image
north-ryde-image
profile-image
request-a-call-image
secure-supported-image
sustainability-image
western-syd-image
snakes_and_Ladders_image
Hyperscale
Eight_Mile
Malaga
North_Ryde
Homebush
blue
green
orange
empty
240521_TechDebtGame.pdf
240201_CS_Health_FMCS_01.pdf
240325_CS_Health_FMCS
Top of Page