Mahdi Zamanighomi | bioRxiv Science

bioRxiv2016

AC-PCA: simultaneous dimension reduction and adjustment for confounding variation

Dimension reduction methods are commonly applied to high-throughput biological datasets. However, the results can be hindered by confounding factors, either biologically or technically originated. In this study, we extend Principal Component Analysis to propose AC-PCA for simultaneous dimension reduction and adjustment for confounding variation. We show that AC-PCA can adjust for a) variations across individual donors present in a human brain exon array dataset, and b) variations of different species in a model organism ENCODE RNA-Seq dataset. Our approach is able to recover the anatomical structure of neocortical regions, and to capture the shared variation among species during embryonic development. For gene selection purposes, we extend AC-PCA with sparsity constraints, and propose and implement an efficient algorithm. The methods developed in this paper can also be applied to more general settings.

Bioinformatics↗

bioRxiv2014

Gene regulatory network inference from perturbed time-series expression data via ordered dynamical expansion of non-steady state actors

The reconstruction of gene regulatory networks from gene expression data has been the subject of intense research activity over the past many decades. A variety of models and methods have been developed to address different aspects of this important problem. However, these techniques are often difficult to scale, are narrowly focused on particular biological and experimental platforms, and require experimental data that are typically unavailable and difficult to ascertain. The more recent availability of higher-throughput sequencing platforms, combined with more precise modes of genetic perturbation, presents an opportunity to formulate more robust and comprehensive approaches to gene network inference. Here, we propose a step-wise framework for identifying gene-gene regulatory interactions that expand from a known point of genetic or chemical perturbation using time series gene expression data. This novel approach sequentially identifies non-steady state genes post-perturbation and incorporates them into a growing series of low-complexity optimization problems. The governing ordinary differential equations of this model are rooted in the biophysics of stochastic molecular events that underlie gene regulation, delineating roles for both protein and RNA-mediated gene regulation. We show the successful application of our core algorithms for network inference using simulated and real datasets.

Bioinformatics↗