Successful Projects



DeepMalaria successfully performs virtual screening for Plasmodium falciparum. State-of-the-art deep learning techniques such as transfer learning and hyper-parameter optimization through external validation are leveraged to create a practical and accurate pipeline in silico.



Anti Microbial Peptides create an important class of novel drug candidates, while low amount of data are available regarding their sequences. AMPDeep overcomes the challenge of low data and beats the state-of-the-art accuracy in prediction of multiple classes of AMPs.

ECG Arrhythmia Classification


Through deep learning the patterns within the Electrocardiogram signals are learnt and utilized to classify arrhythmia within the patients. This work uses the patterns learnt from the 2D domain (images) to featurize the ECG signals. Results showed an scale-able model which is more accurate than human cardiologists.

Ongoing Projects

Allele Frequency Prediction


Allele frequency tracking after any kind of disturbance such as vaccine introduction is still a big challenge. Transfer Learning Is helping us to make tools better than GWAS related ones to find important SNPs.

Sequence-Based Protein Classification


Would primary structure of proteins contain most of their characteristics?! AI is helping us uncover the real role of AA layout in  proteins. Natural patterns are repeating in DNA and Proteins sequences too!

Drug Design


One important AI's application in drug discovery is expanding chemical space and targeting unknown molecules and mechanisms. GAN related algorithms are helping us comping up with more diverse, complex and unusual molecules.