DeepChainTM lets you create your own machine learning tools, or what we call Apps, to solve particular biology questions. We’re democratising AI protein exploration and design by allowing students and independent researchers to tap into our datasets, pre-trained models and supercomputing power. In this series we shout out to developers who are helping our community further its research by building easy-to-use Apps with DeepChainTM.
Can you tell us a bit about yourselves?
We are MEng Students at Stellenbosch University who love to challenge ourselves at solving complex problems. We created our DeepChainTM App as a solution to the UmojaHack DeepChain Antibody Classification Challenge, where we predicted the binding energy of antibody proteins. We hope our app could help biologists design future vaccines.
What’s the challenge that inspired your app?
Influenza pandemics occur when a new flu virus infects people and spreads globally. In 1918, an estimated 500 million people (one-third of the world’s population) became infected with a novel virus, with at least 50 million people dying as a result. But flu is still a globally dangerous virus today, with new variants identified each year. For example, the recent H3N2 variant, a type of swine flu, has posed dangerous pandemic potential.
We now develop annual vaccines to try and prevent such a catastrophe. However, influenza’s genetic makeup constantly evolves, preventing any single vaccine from being 100% effective. As a result, we still have annual “seasonal flu” epidemics, which have a significant impact in terms of lives, healthcare and economic costs. A co-infection with influenza and COVID-19 could carry even higher risks.
For this reason, we need neutralising antibodies that target new flu variants that might escape current vaccines’ protection. As opposed to a vaccine, these would be a therapeutic option for patients who are already infected. Designing a neutralising antibody targeting the H3N2 influenza virus variant could offer enormous therapeutic potential.
What does your DeepchainTM App do?
Our App estimates how effectively an antibody binds to the influenza virus receptor. Given an antibody protein, the app predicts the binding energy to Influenza’s Hemagglutinin protein, which the virus uses to infect cells. Thus, the app can be used to design antibodies for building therapeutics against the flu. The app itself could be used by a range of influenza researchers, predominantly biologists and biomedical researchers.
|Input||Antibody protein sequences|
|Model||BiLSTM, CNN, MLP|
|Output||Binding energy score|
|Training dataset||40k training samples of antibody proteins each consisting of 221 amino acids, each labelled with a binding energy. Source: “antibodybinding” on bio-datasets or Zindi|
|Accuracy||Root Mean Square Error (RMSE): 2.62225|
How can I access DeepChainTM Apps?
To use this App, you can download it from GitHub or access it through our user-friendly DeepChainTM platform. No coding is needed to leverage the power of this and other apps using DeepChainTM. You simply select the app to be used from the menu, input your antibody amino acid sequence of interest and DeepChainTM will provide you with a predicted binding energy.
Join our community!
Visit DeepChain’s™ GitHub to learn more about the technical aspects of building machine learning models and engage with our community of experts. Producing an App helps showcase your skills and network with like-minded experts. You will also be supporting the worldwide scientific community by building open access Apps to help accelerate protein research.
To learn more about DeepChain™, feel free to send us an email at firstname.lastname@example.org!
If you are a computational biologist passionate about AI, join our team! You can find our job vacancies here.