Vaishnavi Kardale, Bioinformatics Centre, Savitribai Phule Pune University
Machine learning has been useful to us in several ways. It has helped in the development of self-driving cars, google assistant, weather prediction, image recognition, language translation, you-tube recommendation, and much more. Machine learning finds it to be very helpful in biology, health informatics, and medical sciences as well. These days machine learning algorithms are being trained to identify cancer just by analyzing an image. Such kinds of experiments are not just far-fetched and have become a very real possibility.
What is deep learning?
Deep learning is a subset of machine learning. The artificial neural network mimics the working of the brain to process data, create patterns, and make decisions. The artificial neural network is based on interconnecting components called perceptrons. A perceptron is a simplified model of a functioning neuron. It has a way for input (of data) and output (of information). There are at least two layers of a perceptron. In between these two layers, there can be many more layers sandwiched. These sandwiched hidden layers make up a ‘deep’ neural network. Deep learning is very good at finding a pattern and making predictions from analyzing a large amount of data which can be incomprehensible to humans. Convolutional Neural Network is an important part of deep learning that enables a computer to learn and program itself from training data. An attention mechanism was devised to highlight the outstanding part and connect the encoder and decoder.
What is Hepatitis B?
Hepatitis B is the inflammation of the liver caused by the effect of the Hepatitis B virus (HBV). Chronic hepatitis can also lead to liver cancer. The HBV enters a hepatocyte (liver cell) it then transfers its relaxed-circular DNA (rcDNA) to the nucleus of the host cell. rcDNA gets converted to covalently closed circular DNA (cccDNA) in the host nucleus. Then by transcription mRNA is made from cccDNA. Further by reverse transcription, the mRNA gets converted into double-stranded linear DNA. This viral DNA then gets integrated into the host genome. The exact site of DNA insertion is randomly distributed on the whole genome with a handful of hotspots. The hotspots for viral DNA insertions were found at repetitive regions, fragile sites, CpG islands and telomeres. The insertion of the viral DNA in a given gene led to altered expression of the product of that gene.
How was deep learning used to identify HBV insertion sites?
DeepHBV is an attention mechanism model for predicting the HBV integration sites accurately and specifically. Its attention mechanism highlights the positions with potentially important biological meaning. It is the first to use CNN for the prediction of the HBV integration sites. DeepHBV was able to identify novel transcription factor binding sites (TFBS) near HBV integration hotspots. DeepHBV learned genomic features automatically and was trained and tested using HBV integration site data from dsVIS database.
The binding site of AR-full site, Arnt, Atf1, bHLHE40, bHLHE41, CLOCK, BMAL1, E2A, cMyc, COUP-TFII, EBF1, Erra and Foxo3 were highlighted by DeepHBV in dsVIS and VISdb databases, to reveal a novel integration preference for HBV. Many proteins encoded by these genes have been reported to be related to tumour suppressor genes. Some of the proteins like CLOCK and BMAL1 are related to the regulation of circadian rhythm which is closely related to HBV-related disease. The biological significance of these transcription factor binding sites on the human genome should be verified by experimental research. DeepHBV hence can be used to highlight the genomic preference of HBV integration and offers a better view and understanding of the molecular mechanism underlying HBV-related cancer.
Also read: Water in the liquid form found in an Ancient Meteorite
Reference:
- Wu, C., Guo, X., Li, M. et al. DeepHBV: a deep learning model to predict hepatitis B virus (HBV) integration sites. BMC Ecol Evo 21, 138 (2021). https://doi.org/10.1186/s12862-021-01869-8
Author info:
Vaishnavi Kardale is a master’s student at the Bioinformatics Centre, Savitribai Phule University. She is interested in protein folding mechanisms and wants to study them further.
- https://bioxone.in/news/worldnews/the-gene-responsible-for-eye-lens-formation-revealed/
- https://bioxone.in/news/worldnews/comeback-of-tuberculosis-but-its-drug-resistant-now/
- https://bioxone.in/news/worldnews/a-drug-to-reduce-covid-infection-by-99/
- https://bioxone.in/news/worldnews/the-effect-of-stroke-on-muscle-sarcomere/
- https://bioxone.in/news/worldnews/artificial-intelligence-ai-for-efficient-covid-testing/
- The Corrosion Prediction from the Corrosion Product Performance
- Nitrogen Resilience in Waterlogged Soybean plants
- Cell Senescence in Type II Diabetes: Therapeutic Potential
- Transgene-Free Canker-Resistant Citrus sinensis with Cas12/RNP
- AI Literacy in Early Childhood Education: Challenges and Opportunities
2 thoughts on “DeepHBV: A machine learning tool to aid in Hepatitis B integration site detection”