TiSA: TimeSeriesAnalysis-a pipeline for the analysis of longitudinal transcriptomics data

Yohan Lefol; Tom Korfage; Robin Mjelle; Christian Prebensen; Torben Lüders; Bruno Müller; Hans Krokan; Antonio Sarno; Lene Alsøe; CONSORTIUM LEMONAID; Jan-Erik Berdal; Pål Sætrom; Hilde Nilsen; Diana Domanska

doi:10.1093/nargab/lqad020

TiSA: TimeSeriesAnalysis-a pipeline for the analysis of longitudinal transcriptomics data

NAR Genom Bioinform. 2023 Mar 3;5(1):lqad020. doi: 10.1093/nargab/lqad020. eCollection 2023 Mar.

Authors

Yohan Lefol^{1

2}, Tom Korfage³, Robin Mjelle⁴, Christian Prebensen^{1

5}, Torben Lüders^{1

6}, Bruno Müller⁷, Hans Krokan⁴, Antonio Sarno⁴, Lene Alsøe^{1

2}; CONSORTIUM LEMONAID; Jan-Erik Berdal^{1

8}, Pål Sætrom^{4

9

10

11}, Hilde Nilsen^{1

2}, Diana Domanska^{2

12}

Affiliations

¹ Institute of Clinical Medicine, University of Oslo, PO Box 1171, Blindern 0318, Norway.
² Department of Microbiology, University of Oslo, Rikshospitalet, Oslo 0424, Norway.
³ Cytura Therapeutics BV, Kloosterstraat 9, Oss 5349AB, The Netherlands.
⁴ Department of Clinical and Molecular Medicine, Norwegian University of Science and Technology, Erling Skjalgsons gate 1, Trondheim 7491, Norway.
⁵ Department of Infectious Diseases, Oslo University Hospital, Oslo 0424, Norway.
⁶ Department of Clinical Molecular Biology, Akershus University Hotspital, Lørenskog 1478, Norway.
⁷ Microsynth AG, Schützenstrasse 15, Balgach CH-9436, Switzerland.
⁸ Department of Infectious Diseases, Akershus University Hotspital, Lørenskog 1478, Norway.
⁹ Department of Computer and Information Science, Norwegian University of Science and Technology, Sem Sælandsvei 9 Gløshaugen, Trondheim 7491, Norway.
¹⁰ Bioinformatics Core Facility-BioCore, Norwegian University of Science and Technology, Erling Skjalgsons gate 1, Trondheim 7491, Norway.
¹¹ K.G. Jebsen Center for Genetic Epidemiology, Norwegian University of Science and Technology, Håkon Jarls gate 11, Trondheim 7491, Norway.
¹² Department of Pathology, Oslo University Hospital-Rikshospitalet, Sognsvannsveien 20, Oslo 0372, Norway.

Abstract

Improved transcriptomic sequencing technologies now make it possible to perform longitudinal experiments, thus generating a large amount of data. Currently, there are no dedicated or comprehensive methods for the analysis of these experiments. In this article, we describe our TimeSeries Analysis pipeline (TiSA) which combines differential gene expression, clustering based on recursive thresholding, and a functional enrichment analysis. Differential gene expression is performed for both the temporal and conditional axes. Clustering is performed on the identified differentially expressed genes, with each cluster being evaluated using a functional enrichment analysis. We show that TiSA can be used to analyse longitudinal transcriptomic data from both microarrays and RNA-seq, as well as small, large, and/or datasets with missing data points. The tested datasets ranged in complexity, some originating from cell lines while another was from a longitudinal experiment of severity in COVID-19 patients. We have also included custom figures to aid with the biological interpretation of the data, these plots include Principal Component Analyses, Multi Dimensional Scaling plots, functional enrichment dotplots, trajectory plots, and complex heatmaps showing the broad overview of results. To date, TiSA is the first pipeline to provide an easy solution to the analysis of longitudinal transcriptomics experiments.