Integrative, normalization-insusceptible statistical analysis of RNA-Seq data, with improved differential expression and unbiased downstream functional analysis

Author(s):  
Dionysios Fanidis ◽  
Panagiotis Moulos

Abstract The study of differential gene expression patterns through RNA-Seq comprises a routine task in the daily lives of molecular bioscientists, who produce vast amounts of data requiring proper management and analysis. Despite widespread use, there are still no widely accepted golden standards for the normalization and statistical analysis of RNA-Seq data, and critical biases, such as gene lengths and problems in the detection of certain types of molecules, remain largely unaddressed. Stimulated by these unmet needs and the lack of in-depth research into the potential of combinatorial methods to enhance the analysis of differential gene expression, we had previously introduced the PANDORA P-value combination algorithm while presenting evidence for PANDORA’s superior performance in optimizing the tradeoff between precision and sensitivity. In this article, we present the next generation of the algorithm along with a more in-depth investigation of its capabilities to effectively analyze RNA-Seq data. In particular, we show that PANDORA-reported lists of differentially expressed genes are unaffected by biases introduced by different normalization methods, while, at the same time, they comprise a reliable input option for downstream pathway analysis. Additionally, PANDORA outperforms other methods in detecting differential expression patterns in certain transcript types, including long non-coding RNAs.

2019 ◽  
Vol 20 (23) ◽  
pp. 6098 ◽  
Author(s):  
Amarinder Singh Thind ◽  
Kumar Parijat Tripathi ◽  
Mario Rosario Guarracino

The comparison of high throughput gene expression datasets obtained from different experimental conditions is a challenging task. It provides an opportunity to explore the cellular response to various biological events such as disease, environmental conditions, and drugs. There is a need for tools that allow the integration and analysis of such data. We developed the “RankerGUI pipeline”, a user-friendly web application for the biological community. It allows users to use various rank based statistical approaches for the comparison of full differential gene expression profiles between the same or different biological states obtained from different sources. The pipeline modules are an integration of various open-source packages, a few of which are modified for extended functionality. The main modules include rank rank hypergeometric overlap, enriched rank rank hypergeometric overlap and distance calculations. Additionally, preprocessing steps such as merging differential expression profiles of multiple independent studies can be added before running the main modules. Output plots show the strength, pattern, and trends among complete differential expression profiles. In this paper, we describe the various modules and functionalities of the developed pipeline. We also present a case study that demonstrates how the pipeline can be used for the comparison of differential expression profiles obtained from multiple platforms’ data of the Gene Expression Omnibus. Using these comparisons, we investigate gene expression patterns in kidney and lung cancers.


Blood ◽  
2019 ◽  
Vol 134 (Supplement_1) ◽  
pp. 1237-1237 ◽  
Author(s):  
Shalini Sankar ◽  
Miriam Guillen Navarro ◽  
Frida Ponthan ◽  
Simon Bomken ◽  
Sirintra Nakjang ◽  
...  

To identify potential regulators of propagation and self-renewal of Acute Lymphoblastic Leukaemia (ALL), we performed an explorative genome-wide RNAi screen followed by CRISPR ex vivo and in vivo validation screens in the t(4;11)-positive ALL cell line SEM. These screens identified the splicing factor PHF5A as a crucial component of the leukemic program. PHF5A is a subunit of the SF3b protein complex, which directs alternative splicing by binding to the branchpoint of pre-mRNA. Mutations in members of this complex including SF3B1 have been implicated in several haematological malignancies. Functional perturbation experiments demonstrated that PHF5A depletion impairs proliferation, viability and clonogenicity in a range of ALL and AML cell lines strongly suggesting that PHF5A is required for leukemic propagation and self-renewal. To identify genetic programs affected by PHF5A inhibition, we performed RNA-seq followed by analysis of differential gene expression and splicing events. We identified 473 genes with differential expression upon PHF5A knockdown. In addition, we performed in-depth analysis of splicing patterns by examining both differential exon/intron usage and exon junction formation. These analyses demonstrated that loss of PHF5A affects splicing of more than 2500 genes with exon skipping and intron retention being the most frequent splicing events. In order to identify processes and pathways affected by PHF5A, we performed gene set enrichment analysis (GSEA) on both differential expression and splicing. While gene sets associated with RNA processing including splicing, turnover and translation were enriched in both data sets, the differential gene expression signature was also linked to DNA repair processes including base excision, mismatch and homologous recombination repair. In line with these findings, knockdown of either PHF5A or its partner protein SF3B1 induced DNA strand breaks as indicated by comet assay and increased y-H2AX levels. Furthermore, both PHF5A and SF3B1 depletion sensitized ALL cells towards the DNA crosslinking agent mitomycin C. Closer inspection of RNA-seq datasets revealed reduced FANCD2 expression and skipping of exon 22 associated with impaired mono-ubiquitination of the FANCD2 protein as a consequence of PHF5A and SF3B1 knockdown. Furthermore, expression of RAD51, a key component of double strand break repair, also decreased upon PHF5A and SF3B1 knockdown. Notably, in vitro pharmacological inhibition of SF3b complex activity using H3B-8800 (or Pladienolide B) showed a very similar effect on FANCD2 expression, and ubiquitination as well as decrease of RAD51 and an increase in y-H2AX levels on a dose and time-dependent manner. This strongly suggests a mechanistic link between impaired RNA splicing and the repair of DNA double-strand breaks. These combined data show that leukemic cells are highly dependent on a functional SF3b splicing complex. Interference with its function results in DNA damage and also sensitizes towards DNA damaging agents pointing towards a possible benefit of the combined application of inhibitors targeting the SF3b complex with more conventional chemotherapy. Disclosures Ponthan: Epistem Ltd: Employment. Zwaan:Sanofi: Consultancy; Incyte: Consultancy; BMS: Research Funding; Roche: Consultancy; Janssen: Consultancy; Daiichi Sankyo: Consultancy; Servier: Consultancy; Jazz Pharmaceuticals: Other: Travel support; Pfizer: Research Funding; Celgene: Consultancy, Research Funding. Vormoor:Abbvie (uncompensated): Consultancy, Honoraria; Novartis: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Roche/Genentech: Consultancy, Honoraria, Research Funding; AstraZeneca: Research Funding.


2015 ◽  
Author(s):  
Xiaobei Zhou ◽  
Mark D Robinson

A correspondence with respect to: Comprehensive evaluation of differential gene expression analysis methods for RNA-seq data Rapaport F, Khanin R, Liang Y, Pirun M, Krek A, Zumbo P, Mason CE, Socci ND and Betel D, Genome Biol 2013, 14:R95


2012 ◽  
Vol 30 (15_suppl) ◽  
pp. 10048-10048
Author(s):  
Dale Han ◽  
Gregory C Bloom ◽  
Marilyn M Bui ◽  
Steven Enkemann ◽  
Hideko Yamauchi ◽  
...  

10048 Background: Liposarcoma (LPS) dedifferentiation signifies conversion to a clinically aggressive phenotype, but the biologic processes required for this change have not been determined. We describe differential gene expression patterns between well-differentiated (WD) and dedifferentiated (DD) tumors to determine pathways involved in LPS dedifferentiation. Methods: From 1999 to 2006, 121 fatty tumors were resected at a single institution. Twenty tumors, consisting of atypical lipomatous tumors (ALT), WD LPS or DD LPS, were randomly selected and clinicopathologic characteristics were retrospectively reviewed. Gene expression profiling was performed on extracted RNA using the Affymetrix GeneChip platform. Differentially expressed genes were obtained and gene network analysis was done using GeneGO by MetaCore. Results: Median age was 59 years and 70% of cases were male. WD tumors, consisting of 3 ALT and 6 WD LPS, were compared with 11 DD LPS. After a median follow-up of 64 months, 7 patients had died of whom 6 had DD LPS. DD histology was associated with lower overall survival (p<0.05). Significance Analysis of Microarrays for WD tumors vs. DD LPS using a 0% false discovery rate showed differential expression of 188 genes. Network analysis of genes from WD tumors vs. DD LPS showed significant (p<0.001) differential regulation of glucose-activated transcription factor ChREBP (carbohydrate response element binding protein), a key element involved in lipogenesis, gluconeogenesis and glycolysis. There was also significant differential regulation of insulin signaling, PI3K-dependent and PKA signal transduction pathways and of amino acid, fatty acid and glucose metabolism pathways (p<0.05). These pathways, based on Gene Ontology cellular processes, mapped to gene networks primarily involved in lipid metabolism (p<0.05). Conclusions: Differential expression of genes involved in lipid metabolism networks is seen in DD LPS and changes in lipid metabolism may be associated with dedifferentiation. These differential gene expression patterns may help identify fatty tumors potentially at risk for progressing to a malignant or DD state and provide prognostic factors and therapeutic targets for patients with LPS.


2019 ◽  
Vol 12 (1) ◽  
pp. 11-19 ◽  
Author(s):  
Jun-Young Shin ◽  
Sang-Heon Choi ◽  
Da-Woon Choi ◽  
Ye-Jin An ◽  
Jae-Hyuk Seo ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document