RNAseq DESeq2¶
All the functions that take place within a cell are performed through proteins. These proteins are coded within the DNA (Deoxyribonucleic acid) of the cell. A gene is a sequence of DNA that encodes for a particular protein. In order to make the necessary proteins, the transcriptional machinary of a cell makes special copies of the respective genes that can be translated to protein sequences. These special copies are called messenger RNAs (Ribonucleic acids).
The amount of mRNA produced by a specific gene is used as a surrogate marker for quantification of the gene activity. With Current highthroughput sequencing technologies (such as in this case RNAseq), all (fragmented) RNA molecules from a biological sample are sequenced. These sequences are then matched to the annotated genome sequence to identify to which gene each sequenced fragment belongs. More sequenced fragments mapping to a gene in one samples as compared to another, means that the specific gene had a higher activity.
This workflow uses the exploratory analysis of RNAseq data using many well stablished Bioconductor packages such as DESeq2, Rsamtools, and GenomicAlignments as described in Love MI et al., 2015
![digraph GRAPH_ID {
fontname="sans-serif";
penwidth="0.1";
compound="true";
edge [comment="Wildcard edge",
fontname="sans-serif",
fontsize=10,
colorscheme="blues3",
color=2,
fontcolor=3];
node [fontname="serif",
fontsize=10,
fillcolor="1",
colorscheme="blues4",
color="2",
fontcolor="4",
style="filled"];
subgraph "cluster0" {
label="Read raw data";
edge [comment="subgraph edge wildcard"];
node [comment="subgraph node wildcard"];
"bams" [shape="invhouse",
label="BAM Files"];
"bamfilelist" [label="BamFileList \n (yieldsize=2e6)"];
"gtf" [shape="invhouse",
label="GTF"];
"c0_mtxdb" [label="makeTxDbFromGFF \n (format=GTF) \n\n exonsBy \n (by=gene)"];
"se" [group=g1,
label="summarizeOverlaps()"];
"dds" [colorscheme="ylorrd3",
label="DESeqDataSet \n (design=~cell+dex)"];
"c0_countdata" [label="assay() \n colData()"];
"ddsMat" [shape="box",
label="DESeqDataSetFromMatrix( \n countdata, colData, \n design=~cell+dex)"];
"bams" -> "bamfilelist";
"gtf" -> "c0_mtxdb";
"c0_mtxdb" -> "se";
"bamfilelist" -> "se";
"se" -> "dds";
"se" -> "c0_countdata";
"c0_countdata" -> "ddsMat";
"dds" -> "ddsMat";
}
subgraph "cluster1" {
label="Exploratory analysis";
edge [comment="subgraph edge wildcard"];
node [comment="subgraph node wildcard"];
"c1_dds" [colorscheme="ylorrd3",
label="DESeqDataSet"];
"plot_rlog" [shape="box",
label="plot() \n plotPCA() \n MDS plot"];
"dds_eSF" [label="estimateSizeFactors()"];
"dist_dds_eSF" [label="dist()"];
"pheatmap_dds" [shape="box",
label="pheatmap() \n MDS Plot"];
"dds_rlog" [label="rlog()"];
"dds_cnt_poisdist" [label="PoissonDistance()"];
"pheatmap_cnt_pois" [shape="box",
label="pheatmap() \n MDS plot"];
"c1_dds" -> "dds_eSF";
"dds_eSF" -> "dist_dds_eSF";
"dist_dds_eSF" -> "pheatmap_dds";
"c1_dds" -> "dds_rlog" [ltail="cluster0"];
"dds_rlog" -> "plot_rlog";
"dds_eSF" -> "dds_cnt_poisdist";
"dds_cnt_poisdist" -> "pheatmap_cnt_pois";
}
subgraph "cluster2" {
label="Differential Expression";
edge [comment="subgraph edge wildcard"];
node [comment="subgraph node wildcard"];
"c2_dds" [label="DESeqDataSet",
colorscheme="ylorrd3"];
"c2_rlog" [label="rlog()"];
"c2_assay" [label="↳ assay() \n ↳ rowVars() \n ↳ order() \n ↳ head()"];
"c2_pheatmap" [shape="box",
label="pheatmap()"];
"dds_eSF_deseq" [label="DESeq() \n results() \n quantile() \n cut()"];
"c2_plotCounts" [shape="box",
label="plotCounts()\n hist()\n plotMA() \n barplot()"];
"c2_res_granges" [label="results(format=GRanges) \n mapIds() \n strand()"];
"c2_init_tracks" [label="GenomeAxisTrack() \n AnnotationTrack() \n DataTrack()"];
"c2_plot_track" [shape="box",
label="plotTrack()"];
"c2_counts" [label="counts(normalized=TRUE) \n svaseq(n.sv=2)"];
"c2_design" [label="design(~SV1+SV2+dex) \n DESeq()"];
"c2_stripchart" [shape="box",
label="stripchart()"];
"c2_dds" -> "dds_eSF_deseq";
"dds_eSF_deseq" -> "c2_plotCounts";
"c2_dds" -> "c2_res_granges";
"c2_res_granges" -> "c2_init_tracks";
"c2_init_tracks" -> "c2_plot_track";
"c2_dds" -> "c2_rlog";
"c2_rlog" -> "c2_assay";
"c2_assay" -> "c2_pheatmap";
"c2_dds" -> "c2_counts";
"c2_counts" -> "c2_design";
"c2_counts" -> "c2_stripchart";
}
}](../../_images/graphviz-4ef647356fe9bb23c4c1247bfc6f4b2fa74cc994.png)
Diagram of RNAseq analysis using DESeq2.
Packages and Dependencies¶
There are 17 packages used in this workflow, which depend on 79 additional packages (dependencies).
Used packages:
- Bioconductor: Rsamtools, GenomicFeatures, GenomicAlignments, BiocParallel, DESeq2, genefilter, AnnotationDbi, org.Hs.eg.db, ReportingTools
- CRAN: pheatmap, RColorBrewer, PoiClaClu, ggplot2, Gviz, fission, sva, fission
Package dependencies:
- Bioconductor: BiocGenerics, Biostrings, GenomeInfoDb, XVector, S4Vectors, rtracklayer, Biobase, biomaRt, IRanges, zlibbioc, GenomicRanges, geneplotter, ggbio, PFAM.db, limma, edgeR, GSEABase, GOstats, VariantAnnotation, BSgenome, OrganismDbi, biovizBase, GO.db
- CRAN: bitops, RCurl, RSQLite, DBI, XML, snow, futile.logger, Rcpp, RcppArmadillo, Hmisc, locfit, plyr, scales, reshape2, gtable, digest, MASS, proto, Formula, lattice, gridExtra, nnet, acepack, latticeExtra, cluster, rpart, foreign, survival, annotate, dichromat, labeling, munsell, stringr, xtable, colorspace, magrittr, stringi, Category, R.utils, hwriter, knitr, GGally, graph, Matrix, RBGL, AnnotationForge, evaluate, markdown, yaml, highr, formatR, reshape, mime, matrixStats, mgcv, nlme
Data¶
RNAseq bam-files from Solaimani Kartalaei P, (2014)