Mining for coexpression across hundreds of datasets using novel rank aggregation and visualization methods

Adler, Priit; Kolde, Raivo; Kull, Meelis; Ткаченко, Александр Георгиевич; Peterson, Hedi; Reimand, Jüri; Vilo, Jaak

doi:10.1186/gb-2009-10-12-r139

Cited by 143 publications

(144 citation statements)

References 54 publications

(55 reference statements)

Supporting

Mentioning

143

Contrasting

Order By: Relevance

“…Each tumor cell line, both untreated and after exposure to 3 Gy, was targeted with pooled siRNAs against each of the selected 89 genes and scored on the basis of cell viability. To identify genes with prosurvival functions common across multiple cell lines tested, we used a rank aggregation approach assuming each cell line was an independent dataset (27,28). With different modes of normalizations and perturbations, LGP2 was invariably the top ranked gene in unirradiated cells (Fig.…”

Section: Resultsmentioning

confidence: 99%

RIG-I–like receptor LGP2 protects tumor cells from ionizing radiation

Widau¹,

Parekh²,

Ranck³

et al. 2014

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

An siRNA screen targeting 89 IFN stimulated genes in 14 different cancer cell lines pointed to the RIG-I (retinoic acid inducible gene I)-like receptor Laboratory of Genetics and Physiology 2 (LGP2) as playing a key role in conferring tumor cell survival following cytotoxic stress induced by ionizing radiation (IR). Studies on the role of LGP2 revealed the following: (i) Depletion of LGP2 in three cancer cell lines resulted in a significant increase in cell death following IR, (ii) ectopic expression of LGP2 in cells increased resistance to IR, and (iii) IR enhanced LGP2 expression in three cell lines tested. Studies designed to define the mechanism by which LGP2 acts point to its role in regulation of IFNβ. Specifically (i) suppression of LGP2 leads to enhanced IFNβ, (ii) cytotoxic effects following IR correlated with expression of IFNβ inasmuch as inhibition of IFNβ by neutralizing antibody conferred resistance to cell death, and (iii) mouse embryonic fibroblasts from IFN receptor 1 knockout mice are radioresistant compared with wild-type mouse embryonic fibroblasts. The role of LGP2 in cancer may be inferred from cumulative data showing elevated levels of LGP2 in cancer cells are associated with more adverse clinical outcomes. Our results indicate that cytotoxic stress exemplified by IR induces IFNβ and enhances the expression of LGP2. Enhanced expression of LGP2 suppresses the IFN stimulated genes associated with cytotoxic stress by turning off the expression of IFNβ.innate immunity | cytoplasmic sensor | interferon beta | DHX58

show abstract

Section: Resultsmentioning

confidence: 99%

RIG-I–like receptor LGP2 protects tumor cells from ionizing radiation

Widau¹,

Parekh²,

Ranck³

et al. 2014

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

show abstract

“…Briefly, RHBDD2 co-expressed genes in different tissue localizations (adrenal gland, brain, breast, colon, head and neck, lung, ovary, small intestine, and stomach) were obtained by using the web-based bioinformatics tool Multiexperiment Matrix (http://biit.cs.ut.ee/mem/; Adler et al 2009). For each tissue, we selected the 200 best positive correlated genes (p <0.0001), which were compiled into one Excel pivot table for comparison of overlapping genes between the tissues, selecting those genes that were present in at least four localizations.…”

Section: Bioinformatics and Statistical Analysismentioning

confidence: 99%

Identification of signaling pathways modulated by RHBDD2 in breast cancer cells: a link to the unfolded protein response

Lacunza

Rabassa

Canzoneri

et al. 2014

Cell Stress and Chaperones

View full text Add to dashboard Cite

Rhomboid domain containing 2 (RHBDD2) was previously observed overexpressed and amplified in breast cancer samples. In order to identify biological pathways modulated by RHBDD2, gene expression profiles of RHBDD2 silenced breast cancer cells were analyzed using whole genome human microarray. Among the statistically significant overrepresented biological processes, we found protein metabolism-with the associated ontological terms folding , ubiquitination , and proteosomal degradation-cell death, cell cycle, and oxidative phosphorylation. In addition, we performed an in silico analysis searching for RHBDD2 co-expressed genes in several human tissues. Interestingly, the functional analysis of these genes showed similar results to those obtained with the microarray data, with negative regulation of protein metabolism and oxidative phosphorylation as the most enriched gene ontology terms. These data led us to hypothesize that RHBDD2 might be involved in endoplasmic reticulum (ER) stress response. Thus, we specifically analyzed the unfolding protein response (UPR) of the ER stress process. We used a lentivirus-based approach for stable silencing of RHBDD2 mRNA in the T47D breast cancer cell line, and we examined the transcriptional consequences on UPR genes as well as the phenotypic effects on migration and proliferation processes. By employing dithiothreitol as an UPR inducer, we observed that cells with silenced RHBDD2 showed increased expression of ATF6, IRE1, PERK, CRT, BiP, ATF4, and CHOP (p <0.01). We also observed that RHBDD2 silencing inhibited colony formation and decreased cell migration. Based on these studies, we hypothesize that RHBDD2 overexpression in breast cancer could represent an adaptive phenotype to the stressful tumor microenvironment by modulating the ER stress response.

show abstract

“…These method, named "data merging" and denoted here with the letter (D), were widely used in [15][16][17] to reconstruct large-scale GRNs because of their simplicity. However, since high dimensional data often suffers from unwanted biases, a variety of techniques can be used to correct for Fig.…”

Section: Data Merging -D Methodsmentioning

confidence: 99%

“…In the "data merging" approach, datasets are integrated at the expression level into a unique dataset, from which GRNs are inferred [15][16][17]. However, one of the major problem of this approach is the removal of batch effects.…”

Section: Introductionmentioning

confidence: 99%

Study of Meta-analysis Strategies for Network Inference Using Information-Theoretic Approaches

Pham

Haibe‐Kains

Bellot

et al. 2016

2016 27th International Workshop on Database and Expert Systems Applications (DEXA)

View full text Add to dashboard Cite

Background: Reverse engineeringof gene regulatory networks (GRNs) from gene expression data is a classical challenge in systems biology. Thanks to high-throughput technologies, a massive amount of gene-expression data has been accumulated in the public repositories. Modelling GRNs from multiple experiments (also called integrative analysis) has; therefore, naturally become a standard procedure in modern computational biology. Indeed, such analysis is usually more robust than the traditional approaches, which suffer from experimental biases and the low number of samples by analysing individual datasets. To date, there are mainly two strategies for the problem of interest: the first one ("data merging") merges all datasets together and then infers a GRN whereas the other ("networks ensemble") infers GRNs from every dataset separately and then aggregates them using some ensemble rules (such as ranksum or weightsum). Unfortunately, a thorough comparison of these two approaches is lacking. Results: In this work, we are going to present another meta-analysis approach for inferring GRNs from multiple studies. Our proposed meta-analysis approach, adapted to methods based on pairwise measures such as correlation or mutual information, consists of two steps: aggregating matrices of the pairwise measures from every dataset followed by extracting the network from the meta-matrix. Afterwards, we evaluate the performance of the two commonly used approaches mentioned above and our presented approach with a systematic set of experiments based on in silico benchmarks. Conclusions:We proposed a first systematic evaluation of different strategies for reverse engineering GRNs from multiple datasets. Experiment results strongly suggest that assembling matrices of pairwise dependencies is a better strategy for network inference than the two commonly used ones.

show abstract

Mining for coexpression across hundreds of datasets using novel rank aggregation and visualization methods

Abstract: The MEM web resource allows users to search for co-expressed genes across all microarray datasets in the ArrayExpress database.

Cited by 143 publications

References 54 publications

RIG-I–like receptor LGP2 protects tumor cells from ionizing radiation

RIG-I–like receptor LGP2 protects tumor cells from ionizing radiation

Identification of signaling pathways modulated by RHBDD2 in breast cancer cells: a link to the unfolded protein response

Study of Meta-analysis Strategies for Network Inference Using Information-Theoretic Approaches

Contact Info

Product

Resources

About