Bringing data to life

End-to-end RNA structure prediction platform with advanced deep leaning approach

- aimed at creating better medicines faster and giving new possibilities of treating untreatable disorders

We are currently working on a deep learning solution that can quickly predict RNA structures with atomic accuracy, allowing for applying our method in pharmacy.

While ∼85% of the human genome is transcribed into RNA, only ∼3% of those transcripts code for protein, indicating that the vast majority of RNAs are noncoding.

 

What is more, a large number of newly discovered noncoding RNAs are disease-associated. RNAs are also known to be targets of small molecule drugs or drug candidates.

 

To enable their function, RNAs adopt specific three-dimensional conformations. Therefore, determining and modelling the RNA structure is key to understanding biological processes, disease mechanisms, and RNA therapies.

 

Although experimental techniques allow for the determination of RNA structures, they are tiresome, time-consuming, costly, require specialized equipment, and not always possible. Computational methods for RNA structure prediction do not have these drawbacks, but existing methods fall far short of atomic accuracy, preventing these methods' application.

eBook (9).png

RNA is a unique biomolecule.
It can both store genetic information and perform biological functions in a variety of molecular processes.

is created to improve
the quality of life and health.

MACHINE LEARNING

approach that incorporates physical and biological knowledge about RNA structure into the design of the deep learning algorithm.

RNA STRUCTURE PREDICTION

based on nucleotide sequence only without the need of experimental data thanks to  pretrained DL network.

DEEP LEARNING

techniques using the attention mechanism fuelled by experimental data. To understand results and assure their fidelity.

IDEA

combining the evolutionary and physical approaches for macromolecule structure prediction.

AI4.BIO - InfoShare - branding (6).png

WHY IS IT IMPORTANT?

15%

of the ~20,000 proteins found in humans can be drugged with small molecules

85%

of the human genome is transcribed into RNA, but only ∼3% of those transcripts code for protein, indicating that the vast majority of RNAs are noncoding 

$2,6B

is the cost of developing one new drug. Computational methods can significantly reduce this cost

38%

 the revenue of global computational biology market was drug discovery and disease modeling applications in 2018

$1.27T

this is how the total global pharmaceutical market was valued at the end of 2020

WHERE COMPUTER SCIENCE MEETS LIFE

Our mission is to enhance more areas of biology & medicine through intensifying the usage of AI in:

  • Biomedical text mining

  • Image analysis

  • Analysis of multi-omics data