Analysis Report
WANG Ziyi
Date: 19 5月 2022
0 Description of Workflow
Previous studies suggested that de novo assembly is beneficial for significantly differentially expressed genes (DEG) even when a reference genome is available. [1] Moreover, it is now widely recognized that performing transcript quantification and afterwards obtaining the gene expression by adding together the expression from the individual transcripts will result in improved gene-level analysis. [2, 3] Therefore, the current analysis for this bulk RNA-seq data set is based on the transcript-level with de novo transcriptome assembly. The current analysis protocol was modified from a guideline attached to a peer-reviewed bioinformatic software called IsoformSwitchAnalyzeR [4] (Please Click HERE to check the origianl protocol). Briefly, the current protocol can be described as a 6-step process:
- After QC and potential quality trimming, the reads were mapped to the genome with the STAR aligner.
- Then the de novo transcriptome assembly was performed with StringTie software and was guided by existing annoation from ENSEMBL database.
- The third step is to merge the assembled transcripts from all individual StringTie runs into one combined reference transcriptome with StringTie software. The merged GTF file was further annotated with SQANTI3 software, which is designed to annotate the data generated from oxford nanopore sequencing. Meanwhile, the Circular RNA was detected with CIRCexplorer2 software [5] from the mapping results created in Step 1 based on the merged GTF file.
- To make the analysis of the individual samples comparable, all samples were re-quantified using the merged GTF file created in Step 3 with a quasi-mapping aligner called Salmon.
- Applying a “rescue” algorithm embbed in the IsoformSwitchAnalyzeR [4] software to the merged GTF file, to ensures that all novel isoforms are assigned to genes and that “reference gene_id” and “reference gene_names” are used instead of “Stringtie gene_is” for all already annotated genes.The “rescued” GTF file was annotated by SQANTI3 software again and the still remained novel genes (“Stringtie genes” not overlapping any “Refrence genes”) was removed for the downstream analysis.
- Finally, after correction of batch effect by RUVSeq software, the significant differentially-expressed genes (DEGs) were detected by DESeq2 software; the significant differential exon usage (DEU) were detected by DEXseq software; the analysis of isoform switches with functional consequences and the associated alternative splicing was performed using IsoformSwitchAnalyzeR software; the significant differentially-expressed circular RNAs were detected by circRNAprofiler software.
The source code is avaible on the Supplementary section to reproduce this analysis.
In addition, other data set was generated in order to use the leafcutter software to check the alternative splicing at nucleotide level; to use the Ularcirc software to visulize, to predict ORF, and to estimate the potenial biological function of detected circular RNAs.
The above workflow is developed and mantained by Bioinformatics Study Group in Okayama University (BSGOU). BSGOU is an international academic community committed to advancing the digital transformation of biological and biomedical research. We unite students, researchers, clinicians, and engineers to collaboratively explore how high-throughput data and integrative computation can drive new theories, models, and discoveries in life sciences. For more information in details, please visit the homepage (https://labonom.github.io/) of BSGOU.
1 Sample Sequencing Statistics
1.1 Reference
- Genome sequence reference:
- Coding sequence reference:
- Non-coding RNA reference:
- Annotation of features:
1.2 Quality Control Report
Click HERE to check the FastQC results.
Click HERE to inspect the de novo transcriptome assembly results.
1.3 Quasi-Mapping by Salmon software
*Mapping Rate (%): the number of reads mapped on genome features (i.e. a genome region contains genes)
Automatically detected most likely library type as ISR.Click HERE to learn more detials about the fragment library types.
1.4 Correction of Batch Effect by RUVSeq
Before Correction
RLE plot
RLE
plot: Relative log expression plots.
PCA plot
After Correction
RLE plot
RLE plot:
Relative log expression plots.
PCA plot
2 Analysis
2.1 Differential Gene Expression Analysis
Click HERE to check the read counts, TPM, and feature length of all mapped genes across all samples in a Microsoft .excel file. Click HERE to check all parameters of DESeq2 model for all detected genes of all samples in a Microsoft .excel file. Click HERE to check the DESeq2 model estimated mean gene expression level of each condition (the intercept of a linear model) and its standard errors. For more detials of the DESeq2 model, please Click HERE.
| Overview of Differentially Expressed Genes (DEGs) | ||||||||||||||
| Sig. AS | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Sig. DEGs | in sig. DEGs | |||||||||||||
| All Dectected Genes |
Non Sig. DEGs |
Up- DEGs |
Down- DEGs |
Genes with Sig. AS |
in Non Sig. DEGs |
in Up- DEGs |
in Down- DEGs |
|||||||
| HEK293_OSMI2_2hA vs HEK293_DMSO_2hA | 16056 | 16054 | 2 | 0 | 467 | 466 | 1 | 0 | ||||||
| HEK293_OSMI2_6hA vs HEK293_DMSO_6hA | 16740 | 16678 | 2 | 60 | 922 | 915 | 0 | 7 | ||||||
| HEK293_TMG_2hB vs HEK293_DMSO_2hB | 15962 | 15962 | 0 | 0 | 536 | 536 | 0 | 0 | ||||||
| HEK293_TMG_6hB vs HEK293_DMSO_6hB | 15529 | 15506 | 1 | 22 | 490 | 490 | 0 | 0 | ||||||
| HEK293_TMG_2hB vs HEK293_OSMI2_2hA | 16486 | 8898 | 4359 | 3229 | 7120 | 3559 | 2439 | 1122 | ||||||
| HEK293_TMG_6hB vs HEK293_OSMI2_6hA | 16118 | 7387 | 4743 | 3988 | 3337 | 1343 | 1447 | 547 | ||||||
| † TPM, Transcripts Per Kilobase Million; DEGs, Differentially-Expressed Genes; Sig., Significant; Up-DEGs, Significantly Up-Regulated Genes; Down-DEGs, Significantly Down-Regulated Genes; AS, Alternative Splicing | ||||||||||||||
2.2 Functional Enrichment Analysis
| Number of Significantlly-Enriched Terms of Functional Enrichment Analysis (FEA) | ||||||||||||||||||||||||
| GSEA | ORA | |||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GO-BP | GO-MF | GO-CC | KEGG | |||||||||||||||||||||
| BP | MF | CC | KEGG | From All Sig. DEGs |
From Up- DEGs |
From Down- DEGs |
From All Sig. DEGs |
From Up- DEGs |
From Down- DEGs |
From All Sig. DEGs |
From Up- DEGs |
From Down- DEGs |
From All Sig. DEGs |
From Up- DEGs |
From Down- DEGs |
|||||||||
| HEK293_OSMI2_2hA vs HEK293_DMSO_2hA | 71 | 32 | 24 | 3 | 23 | 1 | 6 | 5 | 23 | 1 | 6 | 5 | 0 | 0 | 0 | 0 | ||||||||
| HEK293_OSMI2_6hA vs HEK293_DMSO_6hA | 226 | 48 | 32 | 44 | 208 | 41 | 39 | 0 | 0 | 0 | 0 | 0 | 208 | 41 | 39 | 0 | ||||||||
| HEK293_TMG_2hB vs HEK293_DMSO_2hB | 158 | 66 | 43 | 15 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||||||||
| HEK293_TMG_6hB vs HEK293_DMSO_6hB | 88 | 42 | 38 | 14 | 10 | 4 | 4 | 10 | 0 | 0 | 0 | 0 | 10 | 4 | 4 | 10 | ||||||||
| HEK293_TMG_2hB vs HEK293_OSMI2_2hA | 278 | 109 | 80 | 28 | 1281 | 268 | 266 | 102 | 940 | 228 | 175 | 42 | 693 | 136 | 149 | 89 | ||||||||
| HEK293_TMG_6hB vs HEK293_OSMI2_6hA | 294 | 122 | 83 | 14 | 1427 | 275 | 267 | 124 | 1087 | 229 | 202 | 52 | 755 | 126 | 143 | 107 | ||||||||
| † GSEA, Gene Set Enrichment Analysis; ORA, Over Representation Analysis; DEGs, Differentially-Expressed Genes; GO-BP, Gene Ontology Biological Processes; GO-MF, Gene Ontology Molecular Function; GO-CC, Gene Ontology Cellular Component; KEGG, Kyoto Encyclopedia of Genes and Genomes; Sig., Significant; Up-DEGs, Significantly Up-Regulated Genes; Down-DEGs, Significantly Down-Regulated Genes | ||||||||||||||||||||||||
2.3 Analysis of Isoform Switches
Click HERE to download all results of the significantly-switched isoforms analysis (q < 0.05 and dif > 0.05).
Click HERE to download the alternative splicing analysis results for visualization by LeafCutter software.