Open Site Navigation

At the crossroads of Biology, Chemistry, Physics and Artificial Intelligence

We are a deep technology company that creates novel therapeutics using state-of-the-art AI technologies

Projects

Pyxis

The global cancer burden is expected to be 28.4 million cases in 2040, thereby highlighting the need to develop novel anticancer drugs with better therapeutic profiles. Through Pyxis AI-integrated drug discovery platform, we are developing microtubule-targeting drug candidates. Microtubules are subcellular dynamic structures that are required for cell division. to save millions of lives. 

Hydra

We use the Carna AI-integrated drug discovery platform to design novel molecules that have the potential to directly interact with unique higher-order RNA structures of viruses. We are now pursuing these projects internally and through partnerships with academia in order to develop first-in-class antiviral drugs against RNA viruses.

Carna

We use the Carna AI-integrated drug discovery platform to design novel molecules that have the potential to directly interact with unique higher-order RNA structures of viruses. We are now pursuing these projects internally and through partnerships with academia in order to develop first-in-class antiviral drugs against RNA viruses.

The
Numbers

CHEMICAL VALIDITY

100%

Our algorithm is designed to account for chemical rules. All generated molecules are chemically valid.

NOVELTY

95%

The greater majority of our molecules cannot be found in any virtual library.

PATENTABILITY

99%

By exploring untapped chemical space, we are able to generate compounds that nobody has thought of. 

GENERATION RATE

~30K

Unique molecules generated per week per machine.

SYNTHESIZABILITY

30%

Despite being completely de novo, we are able to predict retrosynthetic paths for up to 30% of our top molecules.

MOLECULAR DATA

0

No data, no bias. 

Our platform does not require any reference data.

info@denovosciences.ai

Tel: +374(60)-740-044

Yerevan, Hakob Hakobyan 3

© 2021 Denovo Sciences

  • LinkedIn

Our Team

Hovakim Zakaryan

Co-Founder, Team Lead

Mher Matevosyan

Co-Founder, Lead Data Scientist

Vardan Harutyunyan

Co-Founder, Lead Data Scientist

Irina Tirosyan

Data Scientist

Narek Abelyan

Computational Biologist

Harutyun Sahakyan

Computational Biologist

Hamlet Khachatryan

Computational Chemist

Contact Us

Thanks for submitting!

Voyaging in the Chemical Space

A drastic digitalization of health and pharmaceutical sectors motivates the use of artificial intelligence (AI) to handle large volumes of data, learn and make independent decisions for accomplishing specific tasks like discovery of new drug candidates. The incorporation of AI into the drug development process provides the opportunity to counter the inefficiencies that arise in the classical drug development methods. Our company works on the development and application of new AI technologies that allow us to bring significant improvements to the drug discovery process and create new therapeutics.

While the chemical space of drug-like molecules is estimated to be  >10^60, the number of approved drugs is a few thousands. Chemical space travel algorithms can be used to explore defined regions of chemical space by generating focused virtual libraries for drug discovery. Our AI algorithm explores chemical space for target-specific molecules with better properties than existing drugs. In a few days, our AI-integrated drug discovery platform generates thousands of novel, druggable and synthesizable molecules that match all criteria to be new therapeutic stars.

The current AI technologies in drug discovery require massive datasets of compounds for the learning process. Since an expansive volume of training datasets is not always available for different therapeutic targets, the application and viability of AI algorithms are limited. In contrast, our algorithms learn on biological, chemical and physical simulations, thereby overcoming existing limitations with datasets. This allows us to utilize our AI-integrated drug discovery platform against specific and unique targets, making the undruggable more druggable.

Many diseases like psychiatric disorders or cancer have a multifactorial nature that cannot be cured by the specific modulation of a single target. Therefore, the philosophy of drug discovery is being transformed from “one drug one target” to “one drug multiple targets. Our AI-integrated drug discovery platform opens novel avenues to rationally design the next generation of more effective but less toxic therapeutic agents acting on the multiple targets of the same or different diseases.

Working with Unique Targets

Our AI algorithms have no dependence on training datasets, therefore they can be used to design novel molecules against any therapeutic target even when no datasets are available.

Changing the Paradigm

We create molecules hitting more than one target to achieve a higher therapeutic efficiency.

Pyxis

An AI-integrated drug discovery platform that explores the untapped chemical space to find new therapeutic stars

Targeted AI Drug Discovery

Underneath its core Pyxis utilizes multi-parameter optimization that allows for generation of high-quality molecules satisfying the given parameters.

We use Pyxis for both de-novo drug design, as well as for lead optimization. No matter what target, Pyxis enables efficient discovery of novel molecular structures in a matter of few days.

Technology

Carna

This platform designs novel molecules that have the potential to directly interact with unique higher-order RNA structures

RNA-Targeting Drug Discovery

We use Carna for therapeutically harnessing RNA structures and turning the undruggable into druggable. 

By leveraging our AI platform, Carna opens the door to RNA-targeted drugs and dramatically expands the pool of potential drug targets.

Hydra

Multi-target drug discovery platform for generating molecules hitting more than one target to achieve a higher therapeutic efficiency

Multi-Target Drug Discovery

Building up on the  "one drug multiple targets" philosophy, Hydra utilizes cutting-edge AI technology for generating polypharmacological compounds.

The flexibility of Hydra allows for balancing the activities towards the chosen targets and for designing drugs without any off-target effects. 

S. Sirakanyan, E. Arabyan, A. Hakobyan, T. Hakobyan, G. Chilingaryan, H. Sahakyan, A. Sargsyan, G. Arakelov, K. Nazaryan, R. Izmailyan, L. Abroyan, Z. Karalyan, E. Arakelova, E. Hakobyan, A. Hovakimyan, A. Serobian, M. Neves, J. Ferreira, F. Ferreira & H. Zakaryan

Publications

RNA-targeted small molecules

With the understanding of RNA functions and roles in diseases, as well as the development of synthetic antibiotics that can directly bind to ribosomal RNAs and inhibit bacterial infections, there is growing interest in developing novel RNA-targeted small molecules.

We use the Carna AI-integrated drug discovery platform to design novel molecules that have the potential to directly interact with unique higher-order RNA structures of viruses.

We are now pursuing these projects internally and through partnerships with academia in order to develop first-in-class antiviral drugs against RNA viruses.

Anticancer small molecules

Worldwide, an estimated 19.3 million new cancer cases and 10 million cancer deaths occurred in 2020. The global cancer burden is expected to be 28.4 million cases in 2040, thereby highlighting the need to develop novel anticancer drugs with better therapeutic profiles.

Through a collaborative process, and Pyxis AI-integrated drug discovery platform, we are developing microtubule-targeting drug candidates.

Microtubules are sub-cellular dynamic structures that are required for cell division and considered as a promising therapeutic target to combat cancer.

Projects

Collaborations

Careers

We’re always eager to meet fresh talent, so check out our open positions.

At Denovo Sciences we are working on the development and application of new AI technologies that allow us to bring significant improvements to the drug discovery process and create new therapeutics.

Voyaging in the chemical space


While the chemical space of drug-like molecules is estimated to be  >10^60, the number of approved drugs is several thousands. Chemical space travel algorithms can be used to explore defined regions of chemical space by generating focused virtual libraries for drug discovery. Our AI algorithm explores chemical space for target-specific molecules with better properties than existing drugs. In a few days, our AI-integrated drug discovery platform generates thousands of novel, drug-like and synthesizable molecules that match all criteria to be new therapeutic stars.

Working with unique targets

The current AI technologies in drug discovery require massive datasets of compounds for the learning process. Since an expansive volume of training datasets is not always available for different therapeutic targets, the application and viability of AI algorithms are limited. In contrast, our algorithms learn on biological, chemical and physical simulations, thereby overcoming existing limitations with datasets.

Changing the paradigm

Many diseases like psychiatric disorders or cancer have a multifactorial nature that cannot be cured by the specific modulation of a single target. Therefore, the philosophy of drug discovery is being transformed from “one drug one target” to “one drug multiple targets. Our AI-integrated drug discovery platform opens novel avenues to rationally design the next generation of more effective but less toxic therapeutic agents acting on the multiple targets of the same or different diseases.