Research Interests

Shahid Mukhtar Lab is broadly interested in interdisciplinary research projects at the interface of network biology, bioinformatics, computational modeling, and dynamic interactions in diverse systems including Arabidopsis, humans, and mice. The long-term goal of the Shahid Mukhtar Lab is to understand how macromolecular networks control biological processes and how environmental perturbations in such networks can explain diverse phenotypes. The current projects under investigation include:

Single Cell RNA-Sequencing Analyses

RNA-seq technologies have been widely used for the discovery and relative quantification of transcripts in diverse organisms including animals and plants. Such transcriptomics data provides an opportunity for co-expression network construction to understand the correlation amongst gene expression patterns. Each cell has a unique transcriptome and thus requires an individualized snapshot of its transcription profiles that can be provided by the scRNA-seq sequencing technique. While RNA-seq is only able to highlight expression patterns of large numbers of cells together, scRNA-seq provides a typical window of opportunity to obtain an expression profile at the single-cell level. Currently, we are performing 10X Genomics-based scRNA-seq on diverse cells and nuclei including plant protoplast.

Dynamics of Transcriptional Regulatory Networks

In any eukaryotic cell, thousands of genes and their products orchestrate their transcriptional activities to create cellular functions, phenotypic plasticity, and organismal fecundity. Functional modules embedded within protein–DNA interaction networks execute diverse cellular processes. The dichotomous (deterministic or stochastic) nature of network modules is beneficial to cells or organisms for adaptation to physiological perturbations, environmental cues, or pathological signals. Currently, Shahid Mukhtar Lab is developing a platform by integrating existing and novel computational tools and algorithms that can be exploited to predict, model, and determine the dynamics of host regulatory networks.

Network Biology and Diverse Centrality Measures in Biological SystemsPicture3

Network biology, a branch of systems biology, translates the complexities of molecular interactions into a biological message. The molecules and connections among them are typically called nodes and, edges, respectively. The network-centric structural landscape of networks including co-expression, transcription factors (TF)-target, and protein-protein interactions (interactome) provides a valuable source of information for inferring the functional patterns of genes or their products. For instance, network architectural properties can determine the connectivity and the critical distribution of a particular node within a network. These include degree, the number of connections of a node; and betweenness, the fraction of the shortest paths that pass through a node. In summary, several parameters of these centrality measures may act as indicators of the “most influential nodes” within a network, and prioritize important molecular components in diverse biological systems for testable hypotheses.

Machine Learning, an Analytical Predictive Modeling for Multi-dimensional Biological Datasets 

Machine learning is a component of artificial intelligence that often uses applied statistical and computational techniques to impart the ability to “learn” from data, without being programmed. This is important, as it offers an unguided (i.e., in the absence of user bias) approach to analyzing diverse patterns of data for predictive modeling. Upon learning a set of patterns, the model can predict any range of output. Broadly, machine learning methods are categorized into two groups—namely, supervised and unsupervised learning. In the supervised case, the known features are used to train the machine-learning model prior to making the predictive model. Several sophisticated supervised learning methods include support vector machine (SVM), Multilayer Perceptron (MLP), and Similarity Learning. In contrast, unsupervised learning algorithms do not require known labels or features to infer patterns from a given dataset. In a biological system, machine learning is employed to identify transcriptional start sites, splicing sites, enhancers, promotors, functionally annotate genomes, identify functional redundancy among genes, predict gene expression, metabolism gene prediction, and phenotyping applications using diverse –omics.

Dynamics of Microbial Communities in Response to Abiotic stresses

Picture14Plants are hosts to diverse microbial communities that have been shown to contribute to a variety of plant processes such as growth and disease resistance. Plant microbiomes may colonize external and internal tissues in both above and below-ground environments. While the relationships between roots and soil microbiota have been extensively studied, little is known about the composition and function of the phyllosphere, or leaf, microbiome. The phyllosphere microbiome is much more dynamic than the soil microbiome due to its exposure to rapidly changing environmental factors such as wind, temperature, and moisture. Microbiome composition may additionally be affected by the plant genotype and age. However, these factors may be utilized to explore how plant-derived compounds such as hormones affect microbial diversity and how the plants may recruit certain species of microbes under different conditions including cold, heat, drought, etc. Shahid Mukhtar Lab is conducting a broad-scale study of root and leaf microbiomes in model species.