Amyotrophic Lateral Sclerosis


Searching for disease-modifying treatments for ALS

What is ALS?

ALS is a progressive neurodegenerative disease that causes muscle weakness, paralysis, and ultimately, respiratory failure. In the majority of cases, ALS is fatal, and there is no cure or effective treatment.

Unmet need

Treatments for the current standard of care focus on symptom management and offer little clinical benefit, meaning there is an urgent need to develop effective, disease-modifying treatments for this devastating disease.

Progress so far

We are identifying targets to treat ALS, neurodegenerative diseases and associated mechanisms, using our AI-Target Identification tool and have already progressed two targets to candidate seeking and lead optimization.

ALS affects people of all races and ethnic backgrounds.(1)

It affects as many as 30,000 people in the United States, with 5,000 new cases diagnosed each year.(2)

Increasing incidence and prevalence are reported in different parts of the world.(3)

Only two disease-modifying treatments are currently available(4) but they only extend survival time by 6 months at best.(5)

Our work

AI-augmented drug discovery in ALS

To address the critical unmet need for new treatments, we have focused our research on uncovering potential disease-modifying treatments that could significantly extend patients’ lifespan. This case study will focus on Target 2.

Knowledge: building the right foundations

The process starts with enriching our Biomedical Knowledge Graph - which contains data on all diseases - with ALS specific data. We do this by extracting from unstructured data, such as scientific literature, using NLP algorithms and structured experimentally derived, and omics data generated with postmortem ALS patient tissue samples, which helps us to build a rich and highly detailed picture of the disease.

Target identification: uncovering the best targets

Predictive Target Identification tools then use machine learning to reason across all of this information and predict the most biologically relevant target hypotheses. Our target triage tools also surfaced the accompanying metadata to empower expert drug discoverers to make high quality, data-driven decisions over which targets to progress.

Target validation: a critical step of an effective drug discovery program

In the validation stage, we must prove that the target has potential therapeutic benefit relevant to the disease and that modulating it will have the desired outcome. For a complex disease such as ALS, which has heterogeneous cell types, genetics and patients, we used humanised cellular models - derived from ALS patients - to measure whether the target impacted motor neuron survival. In collaboration with the Sheffield Institute for Translational Neuroscience (SITraN), we showed that compounds modulating both of the targets are neuroprotective against the toxic effects of patient-derived cells, an approach that helped increase scientists' confidence in the relationship between the targets and the disease.

Chemistry: deconstructing chemical complexity to design the right drug

Our AI Chemistry tools combine automation, predictive modelling including feedback loops from our in-house labs, and structure-based drug discovery methods. Our powerful AI models predict on-target and off-target potencies, as well as key ADME properties. Our scientists can then rank and prioritise millions of compounds following complex molecular profiles defined by drug discoverers.

What's next

"We have moved our programmes from AI-identified targets to AI-optimised drug design and we will continue advancing the assets through chemistry and towards the clinic. We are also applying our platform to identify additional targets to treat ALS and other neurodegenerative diseases, such as Parkinson's disease."

Peter Cox  —  VP Drug Discovery, BenevolentAI

(1) National Institute of Neurological Disorders and Stroke. Amyotrophic Lateral Sclerosis (ALS) Fact Sheet. [Accessed April 2021].

(2) John Hopkins Medicine, ALS. Available from:
[Accessed 15/04/21].

(3) Longinetti, E. and Fang, F. (2019). Epidemiology of amyotrophic lateral sclerosis. Current Opinion in Neurology, 32(5), pp.771–776. [Accessed April 2021].

(4) Hardiman, O., Al-Chalabi, A., Chio, A. et al. Amyotrophic lateral sclerosis. Nat Rev Dis Primers 3, 17071 (2017). [Accessed April 2021].

(5) Global Data Report on ALS, September 2020. [Accessed April 2021].