Concordance between ex vivo PBMC and in vivo human infections
confirmed by N-of-1-pathways analysis of single-subject transcriptome
Vincent Gardeux1, Anthony Bosco2, Jianrong Li1, Marilyn J. Halonen3, CARE Network4,
Fernando D. Martinez5,6,*, Yves A. Lussier1,6,7,*
1 Department of Medicine, University of Arizona, Tucson, AZ, USA.
4 The Childhood Asthma Research and Education Network (CARE), University of Arizona, Tucson, AZ, USA
5 Department of Pediatrics, University of Arizona, Tucson, AZ, USA.
6 BIO5 Institute, University of Arizona, Tucson, AZ, USA.
7 UA Cancer Center, University of Arizona, Tucson, AZ, USA.
*Corresponding authors:
Dr. Yves A. Lussier, BIO5 Institute, University of Arizona, 1657 E Helen Street, 251 (P.O. Box 210240), Tucson, AZ, 85721, USA. yves@email.arizona.edu
Dr. Fernando D. Martinez, BIO5 Institute, University of Arizona, 1657 E Helen Street, 251 (P.O. Box 210240), Tucson, AZ, 85721, USA fdmartin@email.arizona.edu
Keywords: personal transcriptome, rhinovirus, PBMC, genomic response, in vivo, ex vivo, viral response
Abstract
Background. Understanding individual patient host-response to viruses is key to designing optimal personalized therapy. Unsurprisingly, in vivo human experimentation to understand individualized dynamic response of the transcriptome to viruses are rarely studied because of the obviously limitations stemming from ethical considerations of the clinical risk. In this rhinovirus study, we first hypothesize that ex vivo human cells response to virus can serve as proxy for otherwise controversial in vivo human experimentation. Of note, comparing the fold change of a few paired measures is the state of the art in human ex vivo assays, which does not scale up to genomics measurements due to excess false positive results. We further hypothesized that the N-of-1-pathways framework, previously validated in cancer, can be effective in understanding the more subtle individual genomic response to viral infection. N-of-1-pathways framework could provide such insight as it is designed to identify deregulated pathways from ontology-anchored gene sets in two paired samples of genome-scale measurements. Finally, we also developed a novel visualization method, similarity Venn Diagram, that provides the similar results between two sets of qualitative measures that can be compared by similarity metrics (e.g. ontology, information theoretic distance, etc).
Method. N-of-1-pathways computes a significance score for a given list of gene sets, using the Ôomics profiles of a mere two samples as input (e.g. normal/tumoral, pre/post-treatment, infected vs non infected cells). We extracted the peripheral blood mononuclear cells (PBMC) of four human subjects, aliquoted in two paired samples one subjected to ex vivo rhinovirus infection. Their deregulated genes and pathways were compared quantitatively and qualitatively as a group to those of 9 human subjects prior and after intranasal inoculation Òin vivoÓ with rhinovirus. Additionally, we developed the Similarity Venn Diagram, an efficient and deceptively simple method for comparing results expressed in an ontology organized as a directed acyclic graph.
Results. We compared the N-of-1-pathways results using two established cohort-level methodologies: GSEA and enrichment of differentially expressed genes. Methodologically, we have extended contingency tables and odds ratio calculation to calculating the significance of Similarity Venn Diagrams. Results are biologically relevant and similar between in vivo and ex vivo studies, both at the genes and enriched pathways levels. Individual patient ROC curves demonstrate that deregulated pathways identified by N-of-1-pathways in PBMC cells of each single subject infected ex vivo recapitulate the biologically relevant pathways observed in vivo in a whole cohort (p=0.004). Further, a principal component analysis of N-of-1-Pathways Scores discriminates asymptotic patients from symptomatic infected patients in vivo (PBMC expression).
Conclusion. There are less than five published transcriptomes of human viral infections in vivo. We show the first evidence that a novel transcriptome analysis of ex vivo essays has the potential to predict individualized response to infectious disease without the clinical risks otherwise associated to in vivo challenges.
Softwares:
http://Lussierlab.org/publications/N-of-1-pathways
http://Lussierlab.org/publications/SimilarityVenn
Supplement data and files:
Supplement File 1 - GSEA
Supplement File 2 - DEG+Enrichment
Supplement File 3 - N-of-1-pathways
Supplement File 4 - ssGSEA
Supplement File 5 - Figure 2
Supplement File 6 - Figure 3
All (Zip file)