Trans-eQTLs are distal to the genes they affect, have low effect sizes and vary across tissues. Because of the issue of low statistical power in identifying them from millions of SNPs which could potentially be true trans-eQTLs, they have been often ignored from studies.
We have come up with a method called Tejaas to find sets of causal trans-eQTLs for different genes in a particular tissue and hypothesize the improvement of single-gene models by inclusion of these trans-eQTLs along with the cis-eQTLs. This work discusses modification of the PrediXcan method and subsequent analysis of model improvement upon including trans-eQTLs, along with addressing issues in the evaluation.

Report

Our collaborators used time-resolved cryo-EM method to obtain a structure of ArfB bound to 70S ribosome with a tRNA carrying a dipeptide in the P site and a short mRNA. We used all-atom molecular dynamics simulations of the complex as well as of free ArfB in explicit solvent to gain further insight into the molecular mechanism of its PTH activity. We were able to observe the dynamic nature of the C-terminal tail of ArfB and the GGQ loop opening up in solution in the absence of the ribosome. With regard to ArfB in the ribosome complex, we think the role of the flexible linker joining the N-terminal globular domain and the C-terminal tail is important in the positioning of NTD for successful catalysis and we tried to look into the mechanism in this project

Report

Ubiquitin has been a protein of interest to study the dynamics at various time scales. Despite being a relatively smaller protein, it shows a range of dynamics and has numerous binding partners. We have also used ubiquitin in this project to test the application of E-CPMG, a method developed previously in the lab that now utilizes the CPMG experiment in its full capacity probing a wide range (4μs-10ms) of timescale. Based on some hints from super-cooled CPMG experiments about putative dynamics inthe ns-μs range, we wanted to characterize this motion. We planned to combine the highpower CPMG experiment along with addition of a viscogen (glycerol) to be able to slowdown motions from the ns-μs into the detectable window of our CPMG experiment

Report

This project involved developing a graphical user interface for a data-based chemical modelling software suite CANDIY developed by Chopra lab. CANDIY as a whole is an integrative platform combining biomolecular modelling, reactivity and machine learning methods. We hoped to extend the accessibilty of the software in the scientific community by developing this easy to use GUI. CANDOCK is a part of this suite and is primarily designed for docking of small molecule ligands with biomolecules. My aim was to give CANDOCK a molecule renderer when we input one to the software (previously only using the command line). This involved integrating CANDOCK with Avogadro, an open-source molecule editor and visualizer.

Poster

My undergraduate projects involved analyzing interactions of various ligands with G-uadruplex DNA/RNA as I worked in the same lab throughout. Primarily, I performed MD simulations and analyzed interaction energies, bond angles and orientations through the time course of the trajectories (as reported below). I also spent some time learning about de novo DNA structure modelling, applications of Machine learning in ligand screening for binding to G4 DNA and performing basic biophysical assays such as CD melting and titration

Report