Studying relative RNA localization From nucleus to the cytosol.

The precise coordination of important biological processes, such as differentiation and development, is highly dependent on the regulation of expression of the genetic information. The flow of the genetic information is tightly regulated on multiple levels. Among them, RNA export to cytosol is an essential step for the production of proteins in eukaryotic cells. Hence, estimating the relative concentration of RNA molecules of a given transcript species in the nucleus and in the cytosol is of major significance as it contributes to the understanding of the dynamics of RNA trafficking between the nucleus and the cytosol. The most efficient way to estimate the levels of RNA species genome-wide is through RNA sequencing (RNAseq). While RNAseq can be performed separately in the nucleus and in the cytosol, because measured transcript levels are relative to the total volume of RNA in these compartments, and because this volume is usually unknown, the transcript levels in the nucleus and in the cytosol cannot be directly compared. Here we show theoretically that if, in addition to nuclear and cytosolic RNA-seq, whole cell RNA-seq is also performed, then accurate estimations of the localization of transcripts can be obtained. Based on this, we designed a method that estimates, first the fraction of the total RNA volume in the cytosol (nucleus), and then, this fraction for every transcript. We evaluate our methodology on simulated data and nuclear and cytosolic single cell data available. Finally, we use our method to investigate the cellular localization of transcripts using bulk RNAseq data from the ENCODE project.

Extensive fragmentation and re-organization of transcription in Systemic Lupus Erythematosus.

Systemic Lupus Erythematosus (SLE) is the prototype of autoimmune diseases, characterized by extensive gene expression perturbations in peripheral blood immune cells. Circumstantial evidence suggests that these perturbations may be due to altered epigenetic profiles and chromatin accessibility but the relationship between transcriptional deregulation and genome organization remains largely unstudied. In this work we propose a genomic approach that leverages patterns of gene coexpression from genome-wide transcriptome profiles in order to identify statistically robust Domains of Co-ordinated gene Expression (DCEs). Application of this method on a large transcriptome profiling dataset of 148 SLE patients and 52 healthy individuals enabled the identification of significant disease-associated alterations in gene co-regulation patterns, which also correlate with SLE activity status. Low disease activity patient genomes are characterized by extensive fragmentation leading to overall fewer DCEs of smaller size. High disease activity genomes display extensive redistribution of co-expression domains with expanded and newly-appearing (emerged) DCEs. The dynamics of domain fragmentation and redistribution are associated with SLE clinical endophenotypes, with genes of the interferon pathway being highly enriched in DCEs that become disrupted and with functions associated to more generalized symptoms, being located in domains that emerge anew in high disease activity genomes. Our results suggest strong links between the SLE phenotype and the underlying genome structure and underline an important role for genome organization in shaping gene expression in SLE.

PyHIST: A Histological Image Segmentation Tool.

The development of increasingly sophisticated methods to acquire high-resolution images has led to the generation of large collections of biomedical imaging data, including images of tissues and organs. Many of the current machine learning methods that aim to extract biological knowledge from histopathological images require several data preprocessing stages, creating an overhead before the proper analysis. Here we present PyHIST (https://github.com/manuel-munoz-aguirre/PyHIST), an easy-to-use, open source whole slide histological image tissue segmentation and preprocessing command-line tool aimed at tile generation for machine learning applications. From a given input image, the PyHIST pipeline i) optionally rescales the image to a different resolution, ii) produces a mask for the input image which separates the background from the tissue, and iii) generates individual image tiles with tissue content.