10 Companies Applying AI for Biologics Discovery

When we hear “AI in drug discovery”, there is a high chance our imagination would draw a small molecule discovery in the first place. Indeed, as it was captured and analyzed in the report The Landscape of Artificial Intelligence (AI) In Pharmaceutical R&D roughly half of more than 300 analyzed AI-driven companies focus on small molecules as a therapeutic modality. This biggest category is then followed by smaller groups of biologics, biomarkers and therapies as an R&D focus. The success of small molecule discovery coupled with AI already demonstrates performance with some “AI-inspired” drug candidates entering the clinical trials, like Relay Therapeutics with their precision oncology small molecules in clinical trials for previously ‘undruggable’ targets, or Exscientia with their small molecule candidate to treat obsessive-compulsive disorder, which began human testing in a phase I trial in 2020 in collaboration with Sumitomo Dainippon Pharma.

#advertisement

At the same time, the number of companies using AI for biologics discovery is also continuously growing. In BiopharmaTrend report there are already 65 selected AI-driven companies focusing on biologics. 

According to the definition by FDA, biological drugs include vaccines, blood and its components, somatic cells, gene therapy, recombinant therapeutic proteins and even tissues. Some narrower examples include therapeutic antibodies for cancer treatment or even COVID-19 vaccines, but all biologics share the same composition “requirement” (sugars, proteins, nucleic acids, their combinations or living entities) and the isolation origin (humans, animals, plants, microorganisms).

RELATED: Will Biologics Surpass Small Molecules In The Pharmaceutical Race?

Below we review ten notable companies applying AI for designing biologics drugs in more detail.

AbCellera Biologics 

Canadian company AbCellera, a biotech company that develops antibody therapeutics, was founded in 2012 and closed the IPO round in the end of 2020 with $555.5M gross proceeds. 

AbCellera aims to search, decode, and analyze natural immune systems to find therapy-suitable antibodies, as well as use humanized antibody platforms to access and assess various functional modalities of the potential therapeutic molecules.

They own sequencing technologies with functional data from single B-cells, which enables researchers to increase the number of antibody candidates. Additionally, their proprietary antibody visualization software Celium helps to deal with the obtained multidimensional data.

After selecting the most promising leads by the number of various criteria, computational protein engineering can be used to optimize antibodies, while also considering the natural antibodies diversity.

Asimov 

Founded in 2017, Asimov is a US-based company building tools to program living cells. Their total funding reached $30M from a number of investors,summing up three funding rounds. 

Asimov owns a mammalian cell engineering platform, which can help to accelerate the design and manufacturing for biologics and gene therapies. First, computer-aided design is used along with the simulation and debugging of genetic systems. 

At the same time, ML bridges large-scale datasets with mechanistic models of biology. In general, AI assists with designing and understanding biological complexity.

Asimov’s focus areas are manufacturing of monoclonal antibodies and multifunctional and difficult-to-express molecules; additionally their technology can cater for vector payloads design and capsid-specific optimization.

BigHat Biosciences

California-based BigHat Biosciences is a protein therapeutics company developing an AI-guided antibody design platform called Milliner. They combine machine learning with synthetic biology, integrating the wet lab for high-speed characterization of the antibodies.

According to BigHat, the strength of their technology is a proprietary wet lab, which goes from in silico designs to expressed, purified, and characterized antibodies in days, by constantly building and testing molecules. 

This design-build-test cycle serves as a “food” for ML algorithms, leading to learning out of each experimental platform cycle. In fact, the AI/ML models learn how mutations affect an antibody's molecular properties, for every property measured by the BigHat workcell.

BigHat was founded in 2019 and already raised a total $99.3M funding, with the most recent $75M series B funding round in July 2022. This round was led by Section 32, but also attracted new investors, such as Amgen Ventures, Bristol Myers Squibb and others. 

Creyon Bio

US-based company with a catchy name Creyon Bio applies an engineering approach to creating new oligonucleotide-based medicines (OBMs). The company was founded in 2019 and raised its $40M funding with a single funding round in March 2022 from a group of investors, led by DCVC Bio and Lux Capital. 

Creyon uses a data-first approach to generate datasets, followed by developing ML models to discover precision oligonucleotide-based medicines. The algorithm works by the principle of mapping the sequence and biophysical properties of existing and novel oligonucleotide-based medicines chemistry, which serves as a soil for the predictive models creation. These models are expected to predict the safety and efficiency of the biological drug. 

The company claims to directly engineer the precision OBMs instead of the trial-and-error approach, so the molecules can be tuned with predictable results before they are tested, helping to save time and money. This could be reached by considering not only the nucleotide sequence of the drug, but also linker and sugar variations, “feeding the data” to the AI/ML algorithm.

Envisagenics

Founded in 2014 as a spinout of Cold Spring Harbor Laboratory, Envisagenics is a New York-based company focusing on the discovery of RNA therapeutics. According to their stated mission, they aim to reduce the complexity of biomedical data with the help of AI/ML technologies. 

Just recently, in August 2022, they received a grant from the National Cancer Institute, resulting in the total raised funding of $27.1M.

Their proprietary AI-driven technology, SpliceCore, is a cloud-based platform that is experimentally validated to predict drug targets and biomarkers through splicing discovery from RNA-sequencing data. According to the company, it ensures a higher precision and speed compared to the traditional methods. 

Additionally, SpliceIO is the immunotherapeutic branch of SpliceCore, which helps to identify tumor neoantigens, potentially targetable with a variety of antibody-based modalities.

Evaxion Biotech

Founded in 2008, a Denmark-based Evaxion Biotech is an AI-driven company, devoted to developing vaccines against cancer and infectious diseases. They own a clinical-stage AI-Immunology platform, combining AI technology with their engineering expertise to generate predictive models, helping to identify unique immunotherapies for patients.

In fact, the AI-Immunology platform consists of three unions, which serve different needs in vaccine development. PIONEER helps with the discovery and design of patient-specific neoepitopes used to derive immuno-oncology therapies. In other words, it is aimed to speed up the discovery of tumor-specific antigens -- those which are present only in the tumor tissue as a result of specific mutations or viral infections. 

On the other hand EDEN and RAVEN are geared towards infectious diseases, bacterial and viral correspondingly. They help to discover novel potent T- and B-cell vaccine designs via optimization of their antigenic and structural properties, as well as prepare them for the production. 

Evaxion pursued a $30M IPO in 2021, after which entered the post-IPO equity funding round in June 2022 worth $40M, led by a single investor Lincoln Park Capital Fund.

HiFiBiO

HiFiBiO is a French biotherapeutics company, which was founded in 2013 and currently has a total funding of $179.5M from multiple investors. The company is geared towards fighting cancer and autoimmune diseases mobilizing the human immune system.

They operate proprietary single cell technologies, which are mostly used to discover and develop antibody therapeutics. HiFiBiO’s Drug Intelligence Science Platform (DIS) combines a single cell platform with AI-based data analysis from patient samples, overcoming such challenges as patient heterogeneity and difficulty of identifying predictive biomarkers.

Their pipeline consists of immunomodulatory antibodies targeting immunosuppressive or inflammatory mechanisms. Two of their oncology candidates recently entered the clinical trials, the results of which are expected to be available in 2024.

Juvena Therapeutics

Juvena is a US-based biopharmaceutical company founded in 2017. With their AI-driven platform they aim to unlock the therapeutic potential of stem cell-secreted proteins and use them for treating various diseases. 

According to Juvena, their AI-enhanced biologics drug discovery and development platform assists with identification of lead therapeutic candidates from a pro-regenerative protein library. It is followed by screening, validating, and engineering these candidates into tissue-specific medicines for degenerative diseases.

Curated by Juvena, the protein library is one of the key tools for their AI-enabled biologics discovery, since it is the map of proteins and their diseases-modifying effects across different organs and indications.

Juvena has raised a total of $8.2, getting their most recent grant in November 2021 from California Institute for Regenerative Medicine.

Nuritas 

Ireland-based Nuritas was founded in 2014, specializing in the discovery of natural active ingredients that can improve human health. They combine artificial intelligence and genomics in their platform to uncover bioactive peptides. 

Their peptide portfolio is used to enrich the products with bioactive peptides, where they can address muscle health, anti-inflammation, gut health, skin care, glucose management and food preservation. 

Nuritas Peptide Finder (N𝛑𝛟) is an artificial intelligence platform with wet lab shaped and trained experimental data. The platform also has curated structured and unstructured data allowing it to predict novel peptides with targeted efficacy. Additionally, N𝛑𝛟™ can predict peptide’s cell penetration or bioavailability characteristics. 

Nuritas has raised total $106.5M funding from a number of investors, with the latest Series B round led by Cleveland Avenue.

ProteinQure

ProteinQure is a Toronto-based company, which was founded in 2017 and currently has a total funding of $4.6M. They focus on designing protein therapeutics using computational approaches. 

The focus of the company is the design of antibodies and nanobodies, for which they claim to be able to optimize the molecules when having functional assays with <100 data points. 

ProteinQure’s computational platform includes high-performance computing, molecular simulations and machine learning to design and optimize protein therapeutics. They aim to solve the possible designing issues connected with a larger size of proteins compared to the small molecules. 

On one hand, ProteinQure’s platform might help with predicting the interactions for difficult targets including GPCRs and some intracellular targets, and on the other hand to reduce the off-target effects. 

RELATED: The Explosion of Therapeutic Modalities: Small Molecules, Biologics, and Everything in Between

Conclusion 

Biologics is the growing multidirectional field in drug discovery, which can considerably benefit from AI/ML technologies. Quite many companies already use AI for the various steps of antibody discovery and development, and other branches of biological drugs such as vaccine development, gene therapy and more. 

The adoption of AI in biologics discovery eased the pain of analyzing and structuring big datasets, predicting the molecule’s properties just by its composition and/or structure, overall helping with understanding complex biological systems. This wouldn’t be possible within the same time frame using just human resources, or some might speculate that it wouldn’t be possible at all. 

Topic: AI in Bio