Combining text mining and machine learning techniques aid us in the identification and extraction of key phrases or terms within or between text, or identify patterns or trends among large collections of text data. By comparative analysis of knowledge contained in ayurvedic texts and modern scientific papers, there exists a possibility of finding common information from both literature. These commonalities, if found, could help to lead a healthy life by sustainable use of potential ayurvedic practices that have support from modern findings. 

Complexity of biological system increases according to the number of pathways and networks. Mutations pave way for producing wrong signals in the pathways which in turn disrupt the entire network. Changes in the networks affect protein synthesis of the pathway and it is phenotypically or genotypically expressed, based on the complexity of the mutations. I am investigating the effect of mutations in biological pathways and its implications in disease pathways using a machine learning approach.

The Non Structural Proteins (NSPs) of SARS- COV2 virus play a major role in its survival and virulence capabilities enabling them to perform host immune system hijacking and manipulation. It is crucial to understand their role during infection and virus induced pathogenicity by characterizing their genome, comparative mutational analysis thereby studying the polymorphisms, structural variations, evolutionary patterns and their influence in the emergence of competent future variants. The core aim of my research is seeking for new mutations in the NSPs which inturn facilitates the process of drug designing.

Various properties of viral proteins (electrostatic, Van der Waals, hydrophobicity) will be graphically extracted from 3D structures. 

Host (human) immunoproteins will also be subjected to image extraction from 3D structures. Using the above data binding sites, function and subcellular locations will be studied. 

The study aims to create a potential model that aids in creating vaccines.

Unraveling the secret behind the energy stimulant property of Arogyapacha using NGS

Leptospirosis is one of the neglected diseases caused by the spirochete, Leptospira interrogans. Various immunoinformatics approaches can be employed for identifying the potential vaccine candidate against the disease. The identified candidate can be cloned and expressed in a suitable expression system for further experimental analysis. In vitro and in vivo biochemical and immunoassays could be performed for better understanding of the vaccine candidate interactions and safety.

Bio-Sequences can be converted into images using Chaos Game Representation (CGR). I investigated the possibility of classification of different organisms from their CGR images of bio-sequences using deep learning and transfer learning models. Different data augmentation techniques  were used in the study and identified its positive effect on the dataset. This study generated a novel SARS-COV-2 CGR image dataset and developed a deep learning model to identify SARS-COV-2 variants from its CGR images. 

Epidemiological data of Stroke, 3 of its subtypes and 7 of its comorbid conditions in the last decade (2009 to 2019) from 204 countries were stratified into 8 global regions. Trend analysis and temporal rank analysis of this data revealed distinct profiles for each global region. Stroke has higher mortality in Asian and African regions compared to high socio-economic regions. This distinction holds in the 1000 genome dataset at the population genetic structure level, as well as gene level as observed in genetic risk variants associated with Stroke and its comorbid conditions. While hypertension is the driving force in Asian and African regions, metabolic conditions like diabetes are the driving force in high socio-economic regions. The current narrative of socio-economic condition of a region being the driving force of variation in burden of Stroke comes into scrutiny with these findings. A SEM model of the genetic risk variants will give a deeper insight into the differences observed.

Using in-silico coupled with in-vitro analysis, I have done a systematic study on the plant phytochemicals of Ayurvedic formulation Medhya Rasayana against Alzheimer’s disease target BACE1. Based on the analysis, two naturally occurring compounds, ‘Covolidine’ from the plant Convolvulus pleuricaulis Choisy and ‘N-(4-hydroxyphenyl) phthalimide’ from Glycyrrhiza glabra were identified as potent hits with high affinity and antagonistic activity against BACE1.  Also, the identified compounds satisfied the pharmacological parameters to act as potent drugs.