Scientific Research

********************

I am currently looking for new opportunities! 

********************

1. Early Embryonic Development under Environmental Stress

2. Evolution of Interspecific Aggression

3. Genetic and Neural correlates of Social Decision-Making

4. Statistical Best Practices

5. Spatiotemporal assessment of MeCP2 function

6. Origin of Meaning: Sound-concept correspondence

7. Memory through metamorphosis?

8. Trans-generational Inheritance

9. Evolution of Morality

10. Cloning Hsp60D cDNA in pUAST vectors

____________________________________

1. Early Embryonic Development under Environmental Stress

During my postdoctoral years in the Lockwood Lab at the University of Vermont, I examined cytoskeletal integrity and mitotic wave synchrony during early embryonic development under environmental stress using live confocal imaging (see figure and the video below) and various computational methods.

screen-shot-2016-11-05-at-2-20-18-pm

Mitotic Waves in early embryos:

Feature detection and tracking: 

unsupervised_phase_classification_Tarun_Gupta

(Above: An example of adaptive hysteresis thresholding and selective watershedding segmentation followed by decomposition and unsupervised clustering for classifying dynamic changes in feature shapes and sizes i.e. chromosomes in different mitotic phases in this case; Below: tracking feature displacement through time)

nuclei_trajectories_TG

jeb_v24_i10_oc_rev

Related to this work, we also study adaptive decoupling in thermo-tolerance  in natural populations. This work was recently published here: Lockwood, B. L., Gupta, T., & Scavotto, R. (2017). Disparate patterns of thermal adaptation between life stages in temperate vs. tropical Drosophila melanogaster. Journal of evolutionary biology.

_____________ *** _____________

2. Evolution of Interspecific Aggression

More on this soon. Please contact me directly if you’d like a copy of the latest manuscript submission on this topic.

_____________ *** _____________

3. Genetic and Neural correlates of Social Decision-Making

The fluidity of various environmental and social factors necessitates plasticity in physiological and behavioral systems as well as an ability to respond rapidly and appropriately to various cues and challenges. Within a social context, specificity and appropriateness of a behavior relies on an organism’s ability to reliably discriminate between species and sexes, factor in any prior encounters or social hierarchy of interacting partners, and adjust to the dynamic shifts in sensory and behavioral feedback cues. In addition to these external factors, internal states influenced by genetic and epigenetic landscapes, hormones, neuromodulators, circadian state, sleep homeostasis, mating-drive etc. shape the underlying structural architecture as well as functional mechanics of organismal behavior.

In graduate school, I explored how cell-type specific alterations in the components of chromatin remodeling machinery, specifically methyl-CpG binding proteins, alter amine neuron function and overall organismal behavior in context of aggression and courtship. For this purpose, I used Drosophila as a model system as it allows expression of various transgenic constructs in a targeted, cell-type specific manner in select neuronal populations at specific developmental stages for high-resolution confocal imaging & behavioral analyses.

Tarun_Gupta_Gal80ts_GFP
Temperature-sensitive Gal80ts system for restricting expression of transgenic constructs to specific developmental stages (left: larvae; right: pupae) 

For instance, the image below highlights dopaminergic neurons (green) in the fly brain (blue) expressing human MeCP2 protein (red) involved in remodeling chromatin architecture through nucleosomal array compaction and oligomerization.

TG_hMeCP2-THneurons_20X

These animals can then be used in various behavioral contexts to assay how these manipulations alter behavioral decision-making in contrast to control organisms. Here’s an example of aggressive lunges in 6-day old, socially naive, transgenic control male flies raised at 18C with post-eclosion 2-day heat-shock treatment at 30C followed by aggression analysis at 25C:

Behaviors such as these can be used to explore cellular and molecular processes associated with context-dependent decision-making, learning and memory of dominance hierarchies and group dynamics in general. The intensity of aggression in this video is not typical of wild-type or transgenic control males. Often, after being hit few times, the ‘loser’ male will typically back-off and move away from the territory/resource (food cup). This highlights intra-specific, inter-individual behavioral variability in clonal populations under identical/controlled environmental conditions.

sr_nature-e1447850667743-219x245

Publication: Gupta et al, (2017) Methyl-CpG binding domain proteins inhibit interspecies courtship and promote aggression in Drosophila. Scientific Reports (7):5420. (Click on the cover for a link to the article)

On a different but related note, this article by Björn Brembs uses such inter-individual variability to reject metaphysical postulation of free will and argue for spontaneous unpredictability as an adaptive biological trait.

_____________ *** _____________

4. Statistical Best Practices:

I am very interested in statistical methods and best practices. Case in point, for determining inherent dimensionality of the dataset or for reducing the number of dimensions or features, it’s common practice in a lot of biology circles to directly estimate the covariance matrices of unscaled variables with wildly different units, or going by various “rules of thumb” like – Kaiser’s criterion, or the elbow in the scree plot of eigen values etc. to determine how many components to eventually retain. In these examples, using unscaled variables will, misleadingly, result in disproportionally loading the variables with biggest units of variance, and using rules-of-thumb like the elbow in the scree-plots may lead you to retaining components or factors that are only really representing noise, like in this example below from my iRelocator project data.

Selecting_n_compoents_randomized_simulation_PCA-Tarun

In this example, if you instinctively go by trying to retain PCA components that cumulatively explain >90% of explained variation, you might be tempted to select 8 components out of 14. I had a lower cut-off / threshold of 70% that suggested me to retain 4 components. And, if you were to just go by retaining the elbow on Cattell’s scree-plot on the right (dark cyan line representing Data), you might end up keeping 5 components unless of course, if you were to simulate noise by running PCA in repeated iterations on randomized matrices and averaging the eigenvalues of the components of these random dataset (dark gray line), will you see that none of the components beyond the 4th, actually explain any meaningful variance in the data. This randomized simulation is implemented by Horn’s Parallel analysis but unfortunately, none of the statistical softwares (except R) offer this test. I implemented this in python and as you can imagine, it’s pretty straightforward to implement no matter what language you are working with.

 

_____________ *** _____________

5. Domain-specific assessment of MeCP2 function

TG_MeCP2

Methyl-CpG binding protein 2 (MeCP2) is one of the key chromatin remodeling proteins involved in dynamic, context dependent regulation of gene expression. As a member of the family of methyl-CpG binding (MBD) proteins, it has traditionally been associated with repression of methylated promoters. With few exceptions, MeCP2 is expressed abundantly in all cell types. Alterations in MeCP2 levels through loss-of-function mutations or gene duplication can lead to an array of neurological conditions under the umbrella of MeCP2 spectrum disorders (MSD). Here, we explored domain-specificity of MeCP2 function by using deviations in sleep/activity levels as a proxy readout of domain-specific alterations in circuit output.

gbbcover

Publication: Gupta et al., (2016) Functional conservation of MBD proteins: MeCP2 and Drosophila MBD proteins alter sleep. Genes Brain Behav.  (8):757-774 (Click on the cover for pdf). 

_____________ *** _____________

6. Origin of Meaning: Sound-concept correspondence

TEDxIt’s often claimed that language is arbitrary. That is, the process through which sounds or symbols acquire correspondence to abstract concepts and ideas is arbitrary. How can natural sciences inform the linguistic models of concept acquisition? Is there a more natural relationship between sounds and their meanings? A naturalist’s perspective on the origin of meaning in language : https://tarungupta.org/2015/02/02/tedx/

_____________ *** _____________

7. Memory through metamorphosis?

Can geospatial learning and memories formed during larval stages survive extensive pruning and neuronal remodeling processes during metamorphosis? We are using a multi-solution maze similar to the prototype indicated below to test this idea.

_____________ *** _____________

8. Trans-generational Inheritance

Recent studies provide a conceptual framework that allows for the possibility of direct, stable, transgenerational transmission of acquired traits through germline epigenetic reprogramming. We are currently exploring maternal effects in context of embryonic thermotolerance. More details to follow.

_____________ *** _____________

9. Evolution of Morality

In 2012, I had the opportunity to participate in an interdisciplinary academic exchange on the subject of evolution of morality at Forum Scientiarum at the Eberhard karls Universität Tübingen in Germany. We had a group of philosophers and a group of biologists/scientists from all over the world discussing the evolutionary roots of moral-decision-making by examining relevant behaviors in non-human animals. The discussions were facilitated by  Prof. Frans de Waal and I delivered a talk on the topic of ‘systematizing our bias in value attribution’.

_____________ *** _____________

10. In-frame cloning of Hsp60D cDNA in HM and EGFP tagged pUAST vectors

This work was conducted during the early days of my research as an undergrad in the Cytogenetics laboratory (Prof. SC Lakhotia) at Banaras Hindu University in India.

XT20-amplified-HSP60D-cDNA-TarunGupta_2008At that time, one of the graduate students in the lab was trying to generate monoclonal antibodies for Hsp60D for localization studies. Because of high sequence and structural identity in all Hsp60 member proteins (A, B, C & D), we tried to tag the Hsp60D protein with well characterized epitopes like EGFP or HM allowing direct visualization of the tagged protein in the living or fixed tissue. HM tag is an in-frame fusion of two different polypeptides: a) Hemagglutinin (antigenic glycoprotein found on the surface of the influenza viruses) and b) Myc (a proto-oncogene involved in the regulation of gene expression). Monoclonal antibodies are available commercially for both of these tags.

______________________________________

Advertisements