Solving bioinformatics problems involves working with large multidimensional datasets. NOVEL has always used the most advanced signal processing methods and artificial intelligence since its foundation. To date, we have accumulated extensive experience in using advanced approaches and creating our own algorithms. The company has powerful computing resources to perform such tasks. Our team uses various technologies, including PyTorch, Catalyst, SciKit-learn, XGboost, Luigi and Docker and often participate in various hackathons and competitions to gain experience in solving new complex problems and share the knowledge. In particular, in 2021, our team won the Open Problems in Single Cell Analysis competition in several categories and presented this achievement at the NeurIPS 2021 conference.
For it's customers NOVEL is able to create solutions based on artificial intelligence methods in expedited manner and provide the results through a web application or on-premises. Tune up and refinement of the solutions allows us to achieve the maximum accuracy in solving customer tasks.
The range of tasks that NOVEL is currently working on with the help of ML approaches are listed below
- instrument image processing (microscopes, DNA sequencers);
- processing of medical images (CT, MRI, X-ray) to assist in making a diagnosis;
- processing of satellite images;
- analysis and modification of protein sequences in order to predict and improve the properties of proteins;
- analysis of peptide sequences for solving problems of immunoinformatics;
- analysis of RNA and DNA sequences to predict taxonomy and other properties;
- predicting drug efficiency using tabular gene expression data;
- prediction of economic and sports performance based on tabular data;
- fundamental research in the field of topology.