Integrative analysis pipeline for pooled CRISPR functional genetic screens - MAGeCKFlute

WubingZhang, Feizhen Wu, and Binbin Wang

13 Nov 2018

Abstract

Genome-wide CRISPR (clustered regularly interspaced short palindrome repeats) coupled with nuclease Cas9 (CRISPR/Cas9) screens represent a promising technology to systematically evaluate gene functions. Data analysis for CRISPR/Cas9 screens is a critical process that includes identifying screen hits and exploring biological functions for these hits in downstream analysis. We have previously developed two algorithms, MAGeCK and MAGeCK-VISPR, to analyze CRISPR/Cas9 screen data in various scenarios. These two algorithms allow users to perform quality control, read count generation and normalization, and calculate beta score to evaluate gene selection performance. In downstream analysis, the biological functional analysis is required for understanding biological functions of these identified genes with different screening purposes. Here, We developed MAGeCKFlute for supporting downstream analysis, utilizing the data provided through MAGeCK and MAGeCK-VISPR. MAGeCKFlute provides several strategies to remove potential biases within sgRNA-level read counts and gene-level beta scores. The downstream analysis with the package includes identifying essential, non-essential, and target-associated genes, and performing biological functional category analysis and pathway enrichment analysis of these genes. The package also visualizes genes in the context of pathways to benefit users exploring screening data. Collectively, MAGeCKFlute enables accurate identification of essential, non-essential, and targeted genes, as well as their related biological functions. This vignette explains the use of the package and demonstrates typical workflows. MAGeCKFlute package version: 1.2.1

How to get help for MAGeCKFlute
Input data
Quick start
Section I: Downstream analysis of MAGeCK RRA
- Negative selection and positive selection
Section II: Downstream analysis of MAGeCK MLE
Session info
References

Note: if you use MAGeCKFlute in published research, please cite:

How to get help for MAGeCKFlute

Any and all MAGeCKFlute questions should be posted to the Bioconductor support site, which serves as a searchable knowledge base of questions and answers:

https://support.bioconductor.org

Posting a question and tagging with “MAGeCKFlute” will automatically send an alert to the package authors to respond on the support site. See the first question in the list of Frequently Asked Questions (FAQ) for information about how to construct an informative post.

You should not email your question to the package authors, as we will just reply that the question should be posted to the Bioconductor support site.

Input data

As input, the MAGeCKFlute package expects gene summary file as obtained by running commands mageck test or mageck mle in MAGeCK (Wei Li and Liu. 2014) and MAGeCK-VISPR (Wei Li and Liu. 2015), which are developed by our lab previously, to analyze CRISPR/Cas9 screen data in different scenarios(Tim Wang 2014, Hiroko Koike-Yusa (2014), Ophir Shalem1 (2014), Luke A.Gilbert (2014), Silvana Konermann (2015)). Both algorithms use negative binomial models to model the variances of sgRNAs, and use Robust Rank Aggregation (for MAGeCK) or maximum likelihood framework (for MAGeCK-VISPR) for a robust identification of selected genes.

MAGeCK-MLE can be used to analyze CRISPR screen data from multi-conditioned experiments. MAGeCK-MLE also normalizes the data across multiple samples, making them comparable to each other. The most important ouput of MAGeCK MLE is “gene_summary” file, which includes the beta scores of multiple conditions and the associated statistics. The ‘beta score’ for each gene describes how the gene is selected: a positive beta score indicates a positive selection, and a negative beta score indicates a negative selection.

MAGeCK RRA allows for the comparison between two experimental conditions. It can identify genes and sgRNAs are significantly selected between the two conditions. The most important output of MAGeCK RRA is the file “gene_summary.txt”. MAGeCK RRA will output both the negative score and positive score for each gene. A smaller score indicates higher gene importance. MAGeCK RRA will also output the statistical value for the scores of each gene. Genes that are significantly positively and negatively selected can be identified based on the p-value or FDR.

Quick start

Here we show the most basic steps for integrative analysis pipeline using gene summary file. Before using MAGeCKFlute, analyzing CRISPR/Cas9 screen data using MAGeCK RRA (in MAGeCK (Wei Li and Liu. 2014)) or MAGeCK MLE (in MAGeCK-VISPR (Wei Li and Liu. 2015)) is necessary, which result in the generation of the gene summary file.

To run MAGeCKFlute pipeline, we need gene summary file generated by running MAGeCK RRA or MAGeCK MLE. MAGeCKFlute package provides two example data, one is MLE_Data, and the other is RRA_Data. We will work with them in this document.

Downstream analysis pipeline for MAGeCK MLE

library(MAGeCKFlute)
##Load gene summary data in MAGeCK MLE results
data("MLE_Data")
##Run the downstream analysis pipeline for MAGeCK MLE
FluteMLE(MLE_Data, ctrlname=c("D7_R1", "D7_R2"), treatname=c("PLX7_R1","PLX7_R2"), prefix="BRAF_", organism="hsa")

All pipeline results are written into local directory “./BRAF_Flute_Results/”, and all figures are integrated into file “BRAF_Flute.mle_summary.pdf”.

Downstream analysis pipeline for MAGeCK RRA

##Load gene summary data in MAGeCK RRA results
data("RRA_Data")
##Run the downstream analysis pipeline for MAGeCK RRA
FluteRRA(RRA_Data, prefix="BRAF", organism="hsa")

All pipeline results are written into local directory “./BRAF_Flute_Results/” too, and all figures are integrated into file “BRAF_Flute.rra_summary.pdf”.

Section I: Downstream analysis of MAGeCK RRA

For experiments with two experimental conditions, we recommend using MAGeCK-RRA to identify essential genes from CRISPR/Cas9 knockout screens and tests the statistical significance of each observed change between two states. Gene summary file in MAGeCK-RRA results summarizes the statistical significance of positive selection and negative selection. Use function ‘data’ to load the dataset, and have a look at the file with a text editor to see how it is formatted.

library(MAGeCKFlute)
data("RRA_Data")
head(RRA_Data)

##      id num  neg.score neg.p.value  neg.fdr neg.rank neg.goodsgrna
## 1  CMIP   6 4.8104e-10  2.8420e-07 0.001650        1             5
## 2  MCL1   3 1.9829e-09  2.8420e-07 0.001650        2             3
## 3 ITGB2   4 2.9590e-08  2.8420e-07 0.001650        3             4
## 4 SCN2A   4 1.9322e-07  8.5260e-07 0.003713        4             3
## 5  GRB2   3 2.6728e-06  1.5631e-05 0.048680        5             2
## 6  EGFR   4 3.0225e-06  1.6768e-05 0.048680        6             3
##    neg.lfc pos.score pos.p.value  pos.fdr pos.rank pos.goodsgrna  pos.lfc
## 1 -0.66277   0.76014     0.73823 0.977480    12939             1 -0.66277
## 2 -1.18000   1.00000     1.00000 1.000000    17140             0 -1.18000
## 3 -1.20780   0.99939     0.99908 0.999967    17115             0 -1.20780
## 4 -0.61912   0.19756     0.27533 0.886655     5329             1 -0.61912
## 5 -0.72466   0.82825     0.80865 0.984303    14036             0 -0.72466
## 6 -0.45355   0.95912     0.94822 0.993803    16331             0 -0.45355

Negative selection and positive selection

Then, extract “neg.fdr” and “pos.fdr” from the gene summary table.

dd.rra = ReadRRA(RRA_Data, organism = "hsa")
head(dd.rra)

##   Official  neg.fdr  pos.fdr ENTREZID
## 1     CMIP 0.001650 0.977480    80790
## 2     MCL1 0.001650 1.000000     4170
## 3    ITGB2 0.001650 0.999967     3689
## 4    SCN2A 0.003713 0.886655     6326
## 5     GRB2 0.048680 0.984303     2885
## 6     EGFR 0.048680 0.993803     1956

We provide a function VolcanoView to visualize top negative and positive selected genes.

RRA_Data$fdr = rowMin(as.matrix(RRA_Data[, c("neg.fdr", "pos.fdr")]))
p1 = VolcanoView(RRA_Data, x = "neg.lfc", y = "fdr", Label = "id")
print(p1)

Take 0.05 as the cutoff, get negative selection and positive selection genes and do enrichment analysis on KEGG pathway and GO BP terms.

Negative selection

universe=dd.rra$ENTREZID
idx = dd.rra$neg.fdr<0.05
geneList= -log10(dd.rra[idx, "neg.fdr"])
names(geneList) = dd.rra$ENTREZID[idx]

kegg.neg = enrichment_analysis(geneList = geneList, universe=universe, 
                             type = "KEGG", plotTitle="KEGG: neg")
bp.neg = enrichment_analysis(geneList = geneList, universe=universe, 
                           type = "GOBP", plotTitle="GOBP: neg")
print(kegg.neg$gridPlot)

print(bp.neg$gridPlot)

Positive selection

universe = dd.rra$ENTREZID
idx = dd.rra$pos.fdr<0.05
geneList = -log10(dd.rra[idx, "pos.fdr"])
names(geneList) = dd.rra$ENTREZID[idx]
kegg.pos=enrichment_analysis(geneList = geneList, universe=universe, 
                             type = "KEGG", plotTitle="KEGG: pos")
bp.pos=enrichment_analysis(geneList = geneList, universe=universe, 
                           type = "GOBP", plotTitle="GOBP: pos")
print(kegg.pos$gridPlot)

print(bp.pos$gridPlot)

Section II: Downstream analysis of MAGeCK MLE

Quality control

** Count summary ** MAGeCK Count in MAGeCK/MAGeCK-VISPR generates a count summary file, which summarizes some basic QC scores at raw count level, including map ratio, Gini index, and NegSelQC. Use function ‘data’ to load the dataset, and have a look at the file with a text editor to see how it is formatted.

data("countsummary")
head(countsummary)

##                                   File    Label    Reads   Mapped
## 1 ../data/GSC_0131_Day23_Rep1.fastq.gz day23_r1 62818064 39992777
## 2  ../data/GSC_0131_Day0_Rep2.fastq.gz  day0_r2 47289074 31709075
## 3  ../data/GSC_0131_Day0_Rep1.fastq.gz  day0_r1 51190401 34729858
## 4 ../data/GSC_0131_Day23_Rep2.fastq.gz day23_r2 58686580 37836392
##   Percentage TotalsgRNAs Zerocounts GiniIndex NegSelQC NegSelQCPval
## 1     0.6366       64076         57   0.08510        0            1
## 2     0.6705       64076         17   0.07496        0            1
## 3     0.6784       64076         14   0.07335        0            1
## 4     0.6447       64076         51   0.08587        0            1
##   NegSelQCPvalPermutation NegSelQCPvalPermutationFDR NegSelQCGene
## 1                       1                          1            0
## 2                       1                          1            0
## 3                       1                          1            0
## 4                       1                          1            0

MapRatesView(countsummary)

IdentBarView(countsummary, x = "Label", y = "GiniIndex", 
             ylab = "Gini index", main = "Evenness of sgRNA reads")

countsummary$Missed = log10(countsummary$Zerocounts)
IdentBarView(countsummary, x = "Label", y = "Missed", fill = "#394E80",
             ylab = "Log10 missed gRNAs", main = "Missed sgRNAs")

** Gene summary ** The gene summary file in MAGeCK-MLE results includes beta scores of all genes in multiple condition samples.

library(MAGeCKFlute)
data("MLE_Data")
head(MLE_Data)

##     Gene sgRNA D7_R1.beta D7_R1.z D7_R1.p.value D7_R1.fdr
## 1 CAMK2B     2  -0.295550 -2.8735      0.459080   0.93481
## 2 ZNF131     3  -0.516610 -4.6916      0.144500   0.77761
## 3  AKAP7     8  -0.065867 -1.0002      0.851310   0.98894
## 4   XPO4     3  -0.414740 -3.7681      0.251170   0.86150
## 5   CDSN     3   0.334000  3.0010      0.043375   0.58305
## 6   SS18     4  -0.587850 -6.7157      0.097669   0.70831
##   D7_R1.wald.p.value D7_R1.wald.fdr D7_R2.beta D7_R2.z D7_R2.p.value
## 1         4.0598e-03     1.3971e-02   -0.29394 -2.8592      0.453500
## 2         2.7100e-06     2.2000e-05   -0.37133 -3.4236      0.304250
## 3         3.1723e-01     4.5758e-01   -0.12551 -1.9012      0.951490
## 4         1.6453e-04     8.7997e-04   -0.30946 -2.8463      0.419990
## 5         2.6910e-03     9.8896e-03    0.33450  3.0107      0.036935
## 6         1.8700e-11     3.4900e-10   -0.48818 -5.6295      0.156380
##   D7_R2.fdr D7_R2.wald.p.value D7_R2.wald.fdr PLX7_R1.beta PLX7_R1.z
## 1   0.92783         4.2468e-03     1.5043e-02    -0.201450   -1.9704
## 2   0.88705         6.1800e-04     2.8788e-03    -0.546960   -5.0925
## 3   0.99669         5.7281e-02     1.2610e-01    -0.068192   -1.0579
## 4   0.92209         4.4235e-03     1.5580e-02    -0.258340   -2.4406
## 5   0.55730         2.6061e-03     9.9515e-03     0.356740    3.2679
## 6   0.80817         1.8100e-08     2.3600e-07    -0.440230   -5.1877
##   PLX7_R1.p.value PLX7_R1.fdr PLX7_R1.wald.p.value PLX7_R1.wald.fdr
## 1        0.647630     0.97100           4.8796e-02       0.10493000
## 2        0.101440     0.74861           3.5300e-07       0.00000361
## 3        0.922660     0.99308           2.9012e-01       0.42105000
## 4        0.496960     0.93711           1.4663e-02       0.03940500
## 5        0.051183     0.65319           1.0834e-03       0.00445010
## 6        0.188270     0.84099           2.1300e-07       0.00000227
##   PLX7_R2.beta PLX7_R2.z PLX7_R2.p.value PLX7_R2.fdr PLX7_R2.wald.p.value
## 1    -0.169390   -1.6398        0.751090     0.98204           1.0104e-01
## 2    -0.329590   -2.9459        0.348730     0.89923           3.2203e-03
## 3    -0.078891   -1.1611        0.947580     0.99823           2.4560e-01
## 4    -0.196110   -1.7581        0.669760     0.97199           7.8736e-02
## 5     0.289550    2.5258        0.080425     0.70649           1.1544e-02
## 6    -0.482790   -5.4036        0.145100     0.79378           6.5300e-08
##   PLX7_R2.wald.fdr
## 1       1.9638e-01
## 2       1.2206e-02
## 3       3.8413e-01
## 4       1.6216e-01
## 5       3.5206e-02
## 6       9.4500e-07

Then, extract beta scores of control and treatment samples from the gene summary table(can be a file path of ‘gene_summary’ or data frame).

data("MLE_Data")
gene_summary = MLE_Data
ctrlname = c("D7_R1", "D7_R2")
treatname = c("PLX7_R1", "PLX7_R2")
#Read beta scores from gene summary table in MAGeCK MLE results
dd=ReadBeta(gene_summary, organism="hsa")
head(dd)

##         Gene EntrezID     D7_R1    D7_R2   PLX7_R1   PLX7_R2
## 816   CAMK2B      816 -0.295550 -0.29394 -0.201450 -0.169390
## 7690  ZNF131     7690 -0.516610 -0.37133 -0.546960 -0.329590
## 9465   AKAP7     9465 -0.065867 -0.12551 -0.068192 -0.078891
## 64328   XPO4    64328 -0.414740 -0.30946 -0.258340 -0.196110
## 1041    CDSN     1041  0.334000  0.33450  0.356740  0.289550
## 6760    SS18     6760 -0.587850 -0.48818 -0.440230 -0.482790

Batch effect removal

Is there batch effects? This is a commonly asked question before perform later analysis. In our package, we provide HeatmapView to ensure whether the batch effect exists in data and use BatchRemove to remove easily if same batch samples cluster together.

##Before batch removal
HeatmapView(dd[,c(ctrlname, treatname)])

batchMat = data.frame(samples = c(ctrlname, treatname), batch = c(1,2,1,2), cov = c(1,1,2,2))
dd1 = BatchRemove(dd[,c(ctrlname, treatname)], batchMat)$data

## Standardizing Data across genes

## After batch removal
HeatmapView(dd1[,c(ctrlname, treatname)])

Normalization of beta scores

It is difficult to control all samples with a consistent cell cycle in a CRISPR screen experiment with multi conditions. Besides, beta score among different states with an inconsistent cell cycle is incomparable. So it is necessary to do the normalization when comparing the beta scores in different conditions. Essential genes are those genes that are indispensable for its survival. The effect generated by knocking out these genes in different cell types is consistent. Based on this, we developed the cell cycle normalization method to shorten the gap of the cell cycle in different conditions. Besides, a previous normalization method called loess normalization is available in this package.(Laurent Gautier 2004)

dd_essential = NormalizeBeta(dd, samples=c(ctrlname, treatname), method="cell_cycle")
head(dd_essential)

##         Gene EntrezID       D7_R1      D7_R2     PLX7_R1    PLX7_R2
## 816   CAMK2B      816 -0.32457335 -0.3341746 -0.26422444 -0.2220722
## 7690  ZNF131     7690 -0.56734169 -0.4221578 -0.71739986 -0.4320962
## 9465   AKAP7     9465 -0.07233521 -0.1426899 -0.08944152 -0.1034270
## 64328   XPO4    64328 -0.45546794 -0.3518190 -0.33884211 -0.2571024
## 1041    CDSN     1041  0.36679918  0.3802865  0.46790483  0.3796033
## 6760    SS18     6760 -0.64557754 -0.5550023 -0.57741140 -0.6329431

#OR
dd_loess = NormalizeBeta(dd, samples=c(ctrlname, treatname), method="loess")
head(dd_loess)

##         Gene EntrezID       D7_R1      D7_R2    PLX7_R1     PLX7_R2
## 816   CAMK2B      816 -0.29280834 -0.2892246 -0.2058229 -0.17247412
## 7690  ZNF131     7690 -0.50187896 -0.3697511 -0.5565985 -0.33626144
## 9465   AKAP7     9465 -0.06430507 -0.1164943 -0.0741203 -0.08354037
## 64328   XPO4    64328 -0.40918172 -0.3067203 -0.2633540 -0.19939397
## 1041    CDSN     1041  0.34659578  0.3653739  0.3281679  0.27465244
## 6760    SS18     6760 -0.56913307 -0.4873820 -0.4493995 -0.49313549

Distribution of all gene beta scores

After normalization, the distribution of beta scores in different conditions should be similar. We can evaluate the distribution of beta scores using the function ‘ViolinView’, ‘DensityView’, and ‘DensityDiffView’.

ViolinView(dd_essential, samples=c(ctrlname, treatname), main="Cell cycle normalized")

DensityView(dd_essential, samples=c(ctrlname, treatname), main="Cell cycle normalized")

DensityDiffView(dd_essential, ctrlname, treatname, main="Cell cycle normalized")

#we can also use the function 'MAView' to evaluate the data quality of normalized
#beta score profile.
MAView(dd_essential, ctrlname, treatname, cex=1, main="Cell cycle normalized")

Estimate cell cycle time by linear fitting

After normalization, the cell cycle time in different condition should be almost consistent. Here we use a linear fitting to estimate the cell cycle time, and use function CellCycleView to view the cell cycle time of all samples.

##Fitting beta score of all genes
CellCycleView(dd_essential, ctrlname, treatname, main="Cell cycle normalized")

Positive selection and negative selection

The function ScatterView can group all genes into three groups, positive selection genes (GroupA), negative selection genes (GroupB), and others, and visualize these three grouped genes in scatter plot. We can also use function RankView to rank the beta score deviation between control and treatment and mark top selected genes in the figure.

p1 = ScatterView(dd_essential, ctrlname, treatname, main="Cell cycle normalized")
print(p1)

## Add column of 'diff'
dd_essential$Control = rowMeans(dd_essential[,ctrlname])
dd_essential$Treatment = rowMeans(dd_essential[,treatname])

rankdata = dd_essential$Treatment - dd_essential$Control
names(rankdata) = dd_essential$Gene
p2 = RankView(rankdata, main="Cell cycle normalized")
print(p2)

Functional analysis of selected genes

For gene set enrichment analysis, we provide three methods in this package, including “ORT”(Over-Representing Test (Guangchuang Yu and He. 2012)), “GSEA”(Gene Set Enrichment Analysis (Aravind Subramanian and Mesirov. 2005)), and “HGT”(hypergeometric test), which can be performed on annotations of Gene ontology(GO) terms (Consortium. 2014), Kyoto encyclopedia of genes and genomes (KEGG) pathways (Minoru Kanehisa 2014), MsigDB gene sets, or custom gene sets. The enrichment analysis can be done easily using function enrichment_analysis, which return a list containing gridPlot (ggplot object) and enrichRes (enrichResult instance). Alternatively, you can do enrichment analysis using the function enrich.ORT for “ORT”, enrich.GSE for GSEA, and enrich.HGT for “HGT”, which return an enrichResult instance. Function EnrichedView and EnrichedGSEView (for enrich.GSE) can be used to generate gridPlot from enrichReseasily, as shown below.

## Get information of positive and negative selection genes
groupAB = p1$data
## select positive selection genes
idx1=groupAB$group=="up"
genes=rownames(groupAB)[idx1]
geneList=groupAB$diff[idx1]
names(geneList)=genes
geneList = sort(geneList, decreasing = TRUE)
universe=rownames(groupAB)
## Do enrichment analysis using HGT method
keggA = enrich.HGT(geneList[1:100], universe, organism = "hsa", limit = c(3, 50))
keggA_grid = EnrichedGSEView(keggA@result, plotTitle = "Positive selection")

## look at the results
head(keggA@result)

##                        ID
## GOBP_0000045 GOBP_0000045
## GOBP_0000050 GOBP_0000050
## GOBP_0000132 GOBP_0000132
## GOBP_0000186 GOBP_0000186
## GOBP_0000188 GOBP_0000188
## GOBP_0000288 GOBP_0000288
##                                                                            Description
## GOBP_0000045                                                    AUTOPHAGOSOME ASSEMBLY
## GOBP_0000050                                                                UREA CYCLE
## GOBP_0000132                              ESTABLISHMENT OF MITOTIC SPINDLE ORIENTATION
## GOBP_0000186                                              ACTIVATION OF MAPKK ACTIVITY
## GOBP_0000188                                             INACTIVATION OF MAPK ACTIVITY
## GOBP_0000288 NUCLEAR-TRANSCRIBED MRNA CATABOLIC PROCESS, DEADENYLATION-DEPENDENT DECAY
##                    NES       pvalue    p.adjust GeneRatio BgRatio geneID
## GOBP_0000045 0.5720577 0.0348542077 0.034854208      1/50   50/54  83460
## GOBP_0000050 0.6566265 0.0014904220 0.003414947      1/10   10/11   1050
## GOBP_0000132 0.7543878 0.0073078443 0.009883388      1/22   22/23  10253
## GOBP_0000186 0.7716724 0.0230459606 0.024610166      1/40   40/43  57551
## GOBP_0000188 0.6786800 0.0086651663 0.011174750      1/24   24/25 161742
## GOBP_0000288 0.6110445 0.0005044887 0.002035202       1/6     6/7  28960
##              geneName Count
## GOBP_0000045     EMC6     1
## GOBP_0000050    CEBPA     1
## GOBP_0000132    SPRY2     1
## GOBP_0000186    TAOK1     1
## GOBP_0000188   SPRED1     1
## GOBP_0000288     DCPS     1

print(keggA_grid)

## Do enrichment analysis using GSEA method
gseA = enrich.GSE(geneList, type = "KEGG", organism = "hsa", pvalueCutoff = 1)
gseA_grid = EnrichedGSEView(gseA@result, plotTitle = "Positive selection")

#should same as
head(gseA@result)

##                          ID                               Description
## KEGG_hsa00190 KEGG_hsa00190                 OXIDATIVE PHOSPHORYLATION
## KEGG_hsa05010 KEGG_hsa05010                         ALZHEIMER DISEASE
## KEGG_hsa05012 KEGG_hsa05012                         PARKINSON DISEASE
## KEGG_hsa04932 KEGG_hsa04932 NON-ALCOHOLIC FATTY LIVER DISEASE (NAFLD)
## KEGG_hsa05169 KEGG_hsa05169              EPSTEIN-BARR VIRUS INFECTION
## KEGG_hsa04260 KEGG_hsa04260                CARDIAC MUSCLE CONTRACTION
##               setSize enrichmentScore      NES      pvalue   p.adjust
## KEGG_hsa00190      45       0.5171211 2.163353 0.001014199 0.05504056
## KEGG_hsa05010      45       0.5232078 2.188817 0.001014199 0.05504056
## KEGG_hsa05012      42       0.5354089 2.220517 0.001017294 0.05504056
## KEGG_hsa04932      39       0.5122680 2.104347 0.001021450 0.05504056
## KEGG_hsa05169      29       0.5261081 2.037966 0.001044932 0.05504056
## KEGG_hsa04260      14       0.6126639 1.954457 0.001158749 0.05504056
##                 qvalues rank                   leading_edge
## KEGG_hsa00190 0.0528552  646 tags=58%, list=23%, signal=45%
## KEGG_hsa05010 0.0528552  646 tags=58%, list=23%, signal=45%
## KEGG_hsa05012 0.0528552  646 tags=60%, list=23%, signal=46%
## KEGG_hsa04932 0.0528552  646 tags=56%, list=23%, signal=44%
## KEGG_hsa05169 0.0528552  662 tags=59%, list=24%, signal=45%
## KEGG_hsa04260 0.0528552  479 tags=64%, list=17%, signal=54%
##                                                                                                                                       core_enrichment
## KEGG_hsa00190 10063/1327/9377/7381/6392/4715/506/4708/4702/1355/4701/10632/1329/100532726/498/7386/29796/513/7385/6389/1337/1340/4716/4719/4709/55967
## KEGG_hsa05010   3028/1327/8883/9377/7381/6392/4715/506/4708/4702/4701/1329/100532726/498/7386/29796/513/7385/6389/1337/1340/2906/4716/4719/4709/55967
## KEGG_hsa05012         1327/9377/7381/6392/4715/506/4708/4702/7332/4701/1329/100532726/498/7386/29796/513/7385/6389/1337/1340/4716/4719/4709/135/55967
## KEGG_hsa04932                    1050/1327/9377/7381/6392/4715/4708/4702/4701/1329/100532726/7186/7386/29796/7385/6389/1337/1340/4716/4719/4709/55967
## KEGG_hsa05169                                                 1460/1387/7979/5608/5434/3312/3305/171568/5704/51728/7186/3106/6693/10621/890/9612/3133
## KEGG_hsa04260                                                                                           1327/9377/7381/1329/7386/29796/7385/1337/1340
##                                                                                                                                                                                                  geneName
## KEGG_hsa00190   COX17/COX4I1/COX5A/UQCRB/SDHD/NDUFB9/ATP5F1B/NDUFB2/NDUFA8/COX15/NDUFA7/ATP5MG/COX5B/NDUFC2-KCTD14/ATP5F1A/UQCRFS1/UQCR10/ATP5F1D/UQCRC2/SDHA/COX6A1/COX6B1/NDUFB10/NDUFS1/NDUFB3/NDUFA12
## KEGG_hsa05010 HSD17B10/COX4I1/NAE1/COX5A/UQCRB/SDHD/NDUFB9/ATP5F1B/NDUFB2/NDUFA8/NDUFA7/COX5B/NDUFC2-KCTD14/ATP5F1A/UQCRFS1/UQCR10/ATP5F1D/UQCRC2/SDHA/COX6A1/COX6B1/GRIN2D/NDUFB10/NDUFS1/NDUFB3/NDUFA12
## KEGG_hsa05012       COX4I1/COX5A/UQCRB/SDHD/NDUFB9/ATP5F1B/NDUFB2/NDUFA8/UBE2L3/NDUFA7/COX5B/NDUFC2-KCTD14/ATP5F1A/UQCRFS1/UQCR10/ATP5F1D/UQCRC2/SDHA/COX6A1/COX6B1/NDUFB10/NDUFS1/NDUFB3/ADORA2A/NDUFA12
## KEGG_hsa04932                                  CEBPA/COX4I1/COX5A/UQCRB/SDHD/NDUFB9/NDUFB2/NDUFA8/NDUFA7/COX5B/NDUFC2-KCTD14/TRAF2/UQCRFS1/UQCR10/UQCRC2/SDHA/COX6A1/COX6B1/NDUFB10/NDUFS1/NDUFB3/NDUFA12
## KEGG_hsa05169                                                                                  CSNK2B/CREBBP/SEM1/MAP2K6/POLR2E/HSPA8/HSPA1L/POLR3H/PSMC4/POLR3K/TRAF2/HLA-B/SPN/POLR3F/CCNA2/NCOR2/HLA-E
## KEGG_hsa04260                                                                                                                                COX4I1/COX5A/UQCRB/COX5B/UQCRFS1/UQCR10/UQCRC2/COX6A1/COX6B1

print(gseA_grid)

For enriched pathways, we can use function KeggPathwayView to visualize the beta score level in control and treatment on pathway map.(Weijun Luo 2013)

genedata = dd_essential[,c("Control","Treatment")]
keggID = gsub("KEGG_", "", gseA@result$ID[1])
#The pathway map will be located on current workspace
KeggPathwayView(gene.data = genedata, pathway.id = keggID, species = "hsa")
##Read the figure into R
pngname=paste0(keggID, ".pathview.multi.png")
grid.arrange(grid::rasterGrob(png::readPNG(pngname)))

file.remove(paste0(keggID, c(".pathview.multi.png", ".png", ".xml")))

## [1] TRUE TRUE TRUE

Identify treatment-associated genes using 9-square model

Considering the difference of beta scores in control and treatment sample, we developed a 9-square model, which group all genes into several subgroups. Among these subgroups, four subgroup genes are treatment-associated, which correspond to specific functions. Group1 and Group3 genes are not selected in the control sample, while they are significantly selected in the treatment sample, so they may be related to drug resistance. Group2 and Group4 genes are selected in control, but they are not selected in treatment, so maybe these genes are associated with drug targets.

p3 = SquareView(dd_essential, label = "Gene", main="Cell cycle normalized")
print(p3)

Functional analysis for treatment-associated genes

Same as the section above. We can do enrichment analysis for treatment-associated genes.

##Get information of treatment-associated genes
Square9 = p3$data
##==select group1 genes in 9-Square
idx=Square9$group=="Group1"
geneList = (Square9$Treatment - Square9$Control)[idx]
names(geneList) = rownames(Square9)[idx]
universe=rownames(Square9)
#====KEGG_enrichment=====
kegg1=enrich.ORT(geneList = geneList, universe = universe, type = "KEGG", limit = c(3, 50))
## look at the results
head(kegg1@result)

##                          ID
## KEGG_hsa00130 KEGG_hsa00130
## KEGG_hsa00030 KEGG_hsa00030
## KEGG_hsa03460 KEGG_hsa03460
## KEGG_hsa00360 KEGG_hsa00360
## KEGG_hsa00100 KEGG_hsa00100
## KEGG_hsa00051 KEGG_hsa00051
##                                                       Description
## KEGG_hsa00130 UBIQUINONE AND OTHER TERPENOID-QUINONE BIOSYNTHESIS
## KEGG_hsa00030                           PENTOSE PHOSPHATE PATHWAY
## KEGG_hsa03460                              FANCONI ANEMIA PATHWAY
## KEGG_hsa00360                            PHENYLALANINE METABOLISM
## KEGG_hsa00100                                STEROID BIOSYNTHESIS
## KEGG_hsa00051                     FRUCTOSE AND MANNOSE METABOLISM
##                     NES     pvalue  p.adjust GeneRatio BgRatio
## KEGG_hsa00130 0.4471863 0.01232870 0.5475148     3/337 10/6556
## KEGG_hsa00030 0.7347985 0.01288270 0.5475148     5/337 28/6556
## KEGG_hsa03460 0.5380828 0.04168216 0.9292608     6/337 50/6556
## KEGG_hsa00360 0.1163067 0.04580930 0.9292608     3/337 16/6556
## KEGG_hsa00100 0.1798251 0.07083623 0.9292608     3/337 19/6556
## KEGG_hsa00051 0.4307888 0.07945428 0.9292608     4/337 32/6556
##                                           geneID
## KEGG_hsa00130                  51805/84274/27235
## KEGG_hsa00030           5226/51071/7086/226/5213
## KEGG_hsa03460 8940/3280/100526739/5429/4292/9894
## KEGG_hsa00360                      5053/314/4128
## KEGG_hsa00100                   8435/120227/1595
## KEGG_hsa00051                 3099/226/5372/5213
##                                            geneName Count
## KEGG_hsa00130                        COQ3/COQ5/COQ2     3
## KEGG_hsa00030               PGD/DERA/TKT/ALDOA/PFKM     5
## KEGG_hsa03460 TOP3B/HES1/CENPS-CORT/POLH/MLH1/TELO2     6
## KEGG_hsa00360                         PAH/AOC2/MAOA     3
## KEGG_hsa00100                  SOAT2/CYP2R1/CYP51A1     3
## KEGG_hsa00051                   HK2/ALDOA/PMM1/PFKM     4

EnrichedGSEView(kegg1@result)

Also, pathway visualization can be done using function KeggPathwayView, the same as the section above.

genedata = dd_essential[, c("Control","Treatment")]
keggID = gsub("KEGG_", "", kegg1@result$ID[1])
KeggPathwayView(gene.data = genedata, pathway.id = keggID, species="hsa")
##Read the figure into R
pngname=paste0(keggID, ".pathview.multi.png")
grid.arrange(grid::rasterGrob(png::readPNG(pngname)))
file.remove(paste0(keggID, c(".pathview.multi.png", ".png", ".xml")))

Session info

sessionInfo()

## R version 3.5.1 Patched (2018-07-12 r74967)
## Platform: x86_64-pc-linux-gnu (64-bit)
## Running under: Ubuntu 16.04.5 LTS
## 
## Matrix products: default
## BLAS: /home/biocbuild/bbs-3.8-bioc/R/lib/libRblas.so
## LAPACK: /home/biocbuild/bbs-3.8-bioc/R/lib/libRlapack.so
## 
## locale:
##  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
##  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=C              
##  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
##  [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
##  [9] LC_ADDRESS=C               LC_TELEPHONE=C            
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       
## 
## attached base packages:
## [1] parallel  stats4    stats     graphics  grDevices utils     datasets 
## [8] methods   base     
## 
## other attached packages:
##  [1] MAGeCKFlute_1.2.1    gridExtra_2.3        pathview_1.22.0     
##  [4] org.Hs.eg.db_3.7.0   AnnotationDbi_1.44.0 IRanges_2.16.0      
##  [7] S4Vectors_0.20.1     Biobase_2.42.0       BiocGenerics_0.28.0 
## [10] ggplot2_3.1.0       
## 
## loaded via a namespace (and not attached):
##   [1] fgsea_1.8.0            colorspace_1.3-2       ggridges_0.5.1        
##   [4] rprojroot_1.3-2        qvalue_2.14.0          XVector_0.22.0        
##   [7] farver_1.0             urltools_1.7.1         ggrepel_0.8.0         
##  [10] bit64_0.9-7            xml2_1.2.0             codetools_0.2-15      
##  [13] splines_3.5.1          GOSemSim_2.8.0         knitr_1.20            
##  [16] jsonlite_1.5           annotate_1.60.0        GO.db_3.7.0           
##  [19] png_0.1-7              pheatmap_1.0.10        graph_1.60.0          
##  [22] ggforce_0.1.3          shiny_1.2.0            compiler_3.5.1        
##  [25] httr_1.3.1             rvcheck_0.1.1          backports_1.1.2       
##  [28] assertthat_0.2.0       Matrix_1.2-15          lazyeval_0.2.1        
##  [31] limma_3.38.2           later_0.7.5            tweenr_1.0.0          
##  [34] htmltools_0.3.6        prettyunits_1.0.2      tools_3.5.1           
##  [37] bindrcpp_0.2.2         igraph_1.2.2           gtable_0.2.0          
##  [40] glue_1.3.0             reshape2_1.4.3         DO.db_2.9             
##  [43] dplyr_0.7.8            fastmatch_1.1-0        Rcpp_1.0.0            
##  [46] enrichplot_1.2.0       Biostrings_2.50.1      nlme_3.1-137          
##  [49] ggraph_1.0.2           stringr_1.3.1          mime_0.6              
##  [52] miniUI_0.1.1.1         clusterProfiler_3.10.0 XML_3.98-1.16         
##  [55] DOSE_3.8.0             europepmc_0.3          zlibbioc_1.28.0       
##  [58] MASS_7.3-51.1          scales_1.0.0           hms_0.4.2             
##  [61] promises_1.0.1         KEGGgraph_1.42.0       RColorBrewer_1.1-2    
##  [64] yaml_2.2.0             memoise_1.1.0          UpSetR_1.3.3          
##  [67] biomaRt_2.38.0         triebeard_0.3.0        ggExtra_0.8           
##  [70] stringi_1.2.4          RSQLite_2.1.1          genefilter_1.64.0     
##  [73] BiocParallel_1.16.0    matrixStats_0.54.0     rlang_0.3.0.1         
##  [76] pkgconfig_2.0.2        bitops_1.0-6           evaluate_0.12         
##  [79] lattice_0.20-38        purrr_0.2.5            bindr_0.1.1           
##  [82] labeling_0.3           cowplot_0.9.3          bit_1.1-14            
##  [85] tidyselect_0.2.5       ggsci_2.9              plyr_1.8.4            
##  [88] magrittr_1.5           R6_2.3.0               DBI_1.0.0             
##  [91] mgcv_1.8-25            pillar_1.3.0           withr_2.1.2           
##  [94] units_0.6-1            survival_2.43-1        KEGGREST_1.22.0       
##  [97] RCurl_1.95-4.11        tibble_1.4.2           crayon_1.3.4          
## [100] rmarkdown_1.10         viridis_0.5.1          progress_1.2.0        
## [103] grid_3.5.1             sva_3.30.0             data.table_1.11.8     
## [106] blob_1.1.1             Rgraphviz_2.26.0       digest_0.6.18         
## [109] xtable_1.8-3           tidyr_0.8.2            httpuv_1.4.5          
## [112] gridGraphics_0.3-0     munsell_0.5.0          viridisLite_0.3.0     
## [115] ggplotify_0.0.3

References

Aravind Subramanian, Vamsi K. Moothaa, Pablo Tamayo, and Jill P. Mesirov. 2005. “Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles.” http://www.pnas.org/content/102/43/15545.full.

Consortium., The Gene Ontology. 2014. “Gene Ontology Consortium: going forward.” https://academic.oup.com/nar/article/43/D1/D1049/2439067.

Guangchuang Yu, Yanyan Han, Li-Gen Wang, and Qing-Yu He. 2012. “clusterProfiler: an R Package for Comparing Biological Themes Among Gene Clusters.” http://online.liebertpub.com/doi/abs/10.1089/omi.2011.0118.

Hiroko Koike-Yusa, E-Pien Tan, Yilong Li. 2014. “Genome-wide recessive genetic screening in mammalian cells with a lentiviral CRISPR-guide RNA library.” http://science.sciencemag.org/content/343/6166/80.long.

Laurent Gautier, Benjamin M. Bolstad, Leslie Cope. 2004. “affy—analysis of Affymetrix GeneChip data at the probe level.” https://academic.oup.com/bioinformatics/article/20/3/307/185980.

Luke A.Gilbert, BrittAdamson, Max A.Horlbeck. 2014. “Genome-Scale CRISPR-Mediated Control of Gene Repression and Activation.” https://linkinghub.elsevier.com/retrieve/pii/S0092-8674(14)01178-7.

Minoru Kanehisa, Yoko Sato, Susumu Goto. 2014. “Data, information, knowledge and principle: back to metabolism in KEGG.” https://academic.oup.com/nar/article-lookup/doi/10.1093/nar/gkt1076.

Ophir Shalem1, *, 2. 2014. “Genome-scale CRISPR-Cas9 knockout screening in human cells.” http://science.sciencemag.org/content/343/6166/84.long.

Silvana Konermann, Alexandro E. Trevino, Mark D. Brigham. 2015. “Genome-scale transcriptional activation by an engineered CRISPR-Cas9 complex.” https://www.nature.com/nature/journal/vnfv/ncurrent/full/nature14136.html.

Tim Wang, David M. Sabatini, Jenny J. Wei1. 2014. “Genetic Screens in Human Cells Using the CRISPR-Cas9 System.” http://science.sciencemag.org/content/343/6166/80.long.

Wei Li, Han Xu, Johannes Köster, and X. Shirley Liu. 2015. “Quality control, modeling, and visualization of CRISPR screens with MAGeCK-VISPR.” https://genomebiology.biomedcentral.com/articles/10.1186/s13059-015-0843-6.

Wei Li, Tengfei Xiao, Han Xu, and X Shirley Liu. 2014. “MAGeCK enables robust identification of essential genes from genome-scale CRISPR/Cas9 knockout screens.” https://genomebiology.biomedcentral.com/articles/10.1186/s13059-014-0554-4.

Weijun Luo, Cory Brouwer. 2013. “Pathview: an R/Bioconductor package for pathway-based data integration and visualization.” https://academic.oup.com/bioinformatics/article-lookup/doi/10.1093/bioinformatics/btt285.