sva + egdeR - differential expression analysis - RNA-seq data
3
0
Entering edit mode
@mrodriguesfernanda-9433
Last seen 7.6 years ago
University of Illinois, Urbana-Champaign

Dear list,
I am performing an RNA-seq analysis for differential gene expression and I have a question regarding the use of the package sva for the estimation of unknown batch effects.

In the sva vignette, it shows examples of using the package for estimation of surrogate variables and then performing DE analysis using the package limma (I am referring to the section 6 of the sva vignette: "Adjusting for surrogate variables using the limma package")

Is that possible to do the same using the package edgeR instead of limma?

Or is sva not compatible with edgeR?

Sorry if this is a dumb question. I am a little new to the bioinformatics world.

 

Thank you!!!
sva edgeR • 6.7k views
ADD COMMENT •
3
Entering edit mode
Jeff Leek ▴ 650
@jeff-leek-5015
Last seen 3.7 years ago
United States

You can estimate the surrogate variables with the svaseq command, then pass them as covariates to the model matrix for edgeR and you should be ok. 

 

Jeff
ADD COMMENT •
1
Entering edit mode

Would it make more sense to apply the svaseq command to normalized cpm data rather than the raw counts?

ADD REPLY •
2
Entering edit mode

In our workflow, we suggest running svaseq on normalized counts, as we estimate size factors on raw counts. If raw counts are provided to svaseq, I think something correlated to sequencing depth should be among the first surrogate variables (I think Jeff confirmed this sometime).

ADD REPLY •
1
Entering edit mode

Yes this is what I've observed previously, for example in these analyses: 

https://github.com/jtleek/svaseq

That if you don't normalize first the first sv is usually representing library size. This is a potential way to normalize but currently unexplored, so I'd stick with normalize, then svaseq like Mike suggests 

 

 

ADD REPLY •
0
Entering edit mode

Thank you all for your quick responses.
So just to clarify and check if I am doing this right. I can normalize my data (using edgeR in my case) and then provide normalized cpm data to svaseq, right?

I am including my codes below, not sure if I am doing this right. If someone could give me some insight that would be nice. My data include 24 samples (12 of each genotype - QQ and qq - and each sex - F and M, having 6 biological replicates for each group - QQ.F, QQ.M, qq.F an qq.M). Not sure if I included the estimated surrogate variables in the model properly.

Also, I tried using the normal sva function instead of svaseq and it gave me very similar results. I heard however that the svaseq function does not really account for TMM normalization factors, which would be important for my analysis. Could someone clarify if this is true? Thank you!

My codes:

## QC, FILTERING AND NORMALIZATION USING EDGER ##

> d <- readDGE(targets, labels = targets$Label, comment.char = "#", header = TRUE,
             columns = c(1,7))

> d <- calcNormFactors(d)

> cpm.values <- cpm(d$counts)

> above1cpm <- rowSums(cpm.values  >= 1)

> d.filt <- d[above1cpm >= 6, , keep.lib.sizes = FALSE]

> dim(d.filt)

# [1] 11679      24

> d.filt <- calcNormFactors(d.filt)

> cpm.values.filt < cpm(d.filt$counts)

> head(cpm.values.filt)

(...)
                                           Samples
                 Tags                     QQ.M.3        QQ.M.4       QQ.M.5       QQ.M.6
  ENSBTAG00000020035   99.33248  182.23360  168.80417    94.63499
  ENSBTAG00000011528   48.37818    55.00343    53.30658    62.32472
  ENSBTAG00000012594   17.57829    32.66077    20.34709    18.87298
  ENSBTAG00000018278 242.30761  244.77718  230.29837  241.66867
  ENSBTAG00000021997   13.03218    25.75558    13.58795    19.61574
  ENSBTAG00000008490   17.31310    22.38235    19.68511    23.73471

> logCPM <- cpm(d.filt, log = T)

> head(logCPM)

# (...)
#                                          Samples
#                  Tags                   QQ.M.5     QQ.M.6
#   ENSBTAG00000020035 7.526067 6.557956
#   ENSBTAG00000011528 5.863254 5.955458
#   ENSBTAG00000012594 4.474122 4.232455
#   ENSBTAG00000018278 7.974202 7.910458
#   ENSBTAG00000021997 3.891917 4.288119
#   ENSBTAG00000008490 4.426424 4.562991

## SVA ##

> Group <- relevel(d.filt$samples$group, ref = "qq.M")
> mod <- model.matrix(~Group)
> colnames(mod)[-1] <- paste0(levels(Group)[-1], "vs", levels(Group)[1])
> head(mod)
#   (Intercept) qq.Fvsqq.M  QQ.Fvsqq.M  QQ.Mvsqq.M
# 1              1                  1                0                   0
# 2              1                  1                0                   0
# 3              1                  1                0                   0
# 4              1                  1                0                   0
# 5              1                  1                0                   0
# 6              1                  1                0                   0

> cont.mod <- makeContrasts(QQ.FvsQQ.M = QQ.Fvsqq.M - QQ.Mvsqq.M,
                          qq.Fvsqq.M = qq.Fvsqq.M,
                          QQ.Fvsqq.F = QQ.Fvsqq.M - qq.Fvsqq.M,
                          QQ.Mvsqq.M = QQ.Mvsqq.M,
                          Interact = (QQ.Fvsqq.M - QQ.Mvsqq.M) - (qq.Fvsqq.M),
                          levels = mod)
> cont.mod
# Contrasts
# Levels                 QQ.FvsQQ.M  qq.Fvsqq.M  QQ.Fvsqq.F  QQ.Mvsqq.M  Interact
#         Intercept                       0                  0                  0                    0             0
#     qq.Fvsqq.M                       0                  1                 -1                    0            -1
#    QQ.Fvsqq.M                       1                  0                  1                    0             1
#   QQ.Mvsqq.M                      -1                  0                  0                    1            -1

> svseq = svaseq(cpm.values.filt, mod, mod[,1])
# Number of significant surrogate variables is:  5
# Iteration (out of 5 ):1  2  3  4  5  

> head(svseq$sv)
                    [,1]                [,2]                [,3]                [,4]               [,5]
[1,]  0.05741827 -0.06602954  0.20009240 -0.02684960  0.23766642
[2,]  0.26314274 -0.66218564  0.19546081  0.38882480 -0.21056924
[3,] -0.06735629  0.01038846  0.08790005  0.02410775  0.34376335
[4,]  0.10120300  0.10301322  0.24170332  0.15549638  0.06207581
[5,] -0.17679364 -0.09238517 -0.03882105  0.15047943  0.23952769
[6,]  0.07944103 -0.01603642 -0.26944731 -0.16374470 -0.23344724

> mod.sv <- cbind(mod, svseq$sv)
> colnamesmod.sv)[5:9] <- paste0("sv", 1:5)

#   (Intercept)  qq.Fvsqq. M  QQ.Fvsqq.M  QQ.Mvsqq.M         sv1                      sv2                sv3              sv4              sv5
# 1           1                        1                  0                     0    0.05741827 -0.06602954  0.20009240 -0.02684960  0.23766642
# 2           1                        1                  0                     0    0.26314274 -0.66218564  0.19546081  0.38882480 -0.21056924
# 3           1                        1                  0                     0   -0.06735629  0.01038846  0.08790005  0.02410775  0.34376335
# 4           1                        1                  0                     0    0.10120300  0.10301322  0.24170332  0.15549638  0.06207581
# 5           1                        1                  0                     0   -0.17679364 -0.09238517 -0.03882105  0.15047943  0.23952769
# 6           1                        1                  0                     0    0.07944103 -0.01603642 -0.26944731 -0.16374470 -0.23344724

> cont.mod.sv <- makeContrasts(QQ.FvsQQ.M = QQ.Fvsqq.M - QQ.Mvsqq.M,
                             qq.Fvsqq.M = qq.Fvsqq.M,
                             QQ.Fvsqq.F = QQ.Fvsqq.M - qq.Fvsqq.M,
                             QQ.Mvsqq.M = QQ.Mvsqq.M,
                             Interact = (QQ.Fvsqq.M - QQ.Mvsqq.M) - (qq.Fvsqq.M),
                             levels = mod.sv)

> cont.mod.sv
# Contrasts
# Levels             QQ.FvsQQ.M   qq.Fvsqq.M   QQ.Fvsqq.F   QQ.Mvsqq.M   Interact
#        Intercept                   0                     0                   0                   0              0
#     qq.Fvsqq.M                  0                     1                  -1                   0             -1
#    QQ.Fvsqq.M                  1                     0                   1                   0              1
#   QQ.Mvsqq.M                 -1                     0                   0                   1             -1
#                  sv1                 0                      0                   0                   0              0
#                  sv2                 0                      0                   0                   0              0
#                  sv3                 0                      0                   0                   0              0
#                  sv4                 0                      0                   0                   0              0
#                  sv5                 0                      0                   0                   0              0

> d.filt.sv <- estimateGLMCommonDisp(d.filt, mod.sv, verbose = TRUE)
# Disp = 0.0268 , BCV = 0.1637    ###not sure if the BCV value is good. I am working on bovine. edgeR recommends
                                                              0.1 for identical strains and 0.4 for humans. Not sure if 0.1637 is okay for bovine.

> d.filt.sv <- estimateGLMTrendedDisp(d.filt, mod.sv)
> d.filt.sv <- estimateGLMTagwiseDisp(d.filt, mod.sv)

> plotBCVd.filt.sv)

> fit.edgeR.sv <- glmFitd.filt.sv, mod.sv)
 

> eR.QQ.FvsQQ.M.sv <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 1])
> eR.qq.Fvsqq.M.sv   <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 2])
> eR.QQ.Fvsqq.F.sv   <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 3])
> eR.QQ.Mvsqq.M.sv <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 4])
> eR.Interact.sv <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 5])

# I am not sure on how to check on sv significance from here or if what I did above is right. Should I have the svs in the model another way? Any insight are welcome! Thank you again

R version 3.2.2 (2015-08-14)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu precise (12.04.5 LTS)

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats4    parallel  stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] sva_3.14.0           genefilter_1.50.0    mgcv_1.8-10          nlme_3.1-122         rtracklayer_1.28.10
 [6] GenomicRanges_1.20.8 GenomeInfoDb_1.4.3   IRanges_2.2.9        S4Vectors_0.6.6      rgl_0.95.1429       
[11] edgeR_3.10.5         limma_3.24.15        affycoretools_1.40.5 affy_1.46.1          Biobase_2.28.0      
[16] BiocGenerics_0.14.0  WGCNA_1.48           RSQLite_1.0.0        DBI_0.3.1            fastcluster_1.1.16  
[21] dynamicTreeCut_1.62

loaded via a namespace (and not attached):
 [1] Category_2.34.2           bitops_1.0-6              matrixStats_0.50.1        doParallel_1.0.10        
 [5] RColorBrewer_1.1-2        gcrma_2.40.0              tools_3.2.2               affyio_1.36.0            
 [9] KernSmooth_2.23-15        rpart_4.1-10              Hmisc_3.17-1              colorspace_1.2-6         
[13] nnet_7.3-11               gridExtra_2.0.0           GGally_0.5.0              DESeq2_1.8.2             
[17] bit_1.1-12                preprocessCore_1.30.0     graph_1.46.0              ggbio_1.16.1             
[21] caTools_1.17.1            scales_0.3.0              RBGL_1.44.0               stringr_1.0.0            
[25] Rsamtools_1.20.5          foreign_0.8-66            R.utils_2.2.0             AnnotationForge_1.10.1   
[29] XVector_0.8.0             dichromat_2.0-0           BSgenome_1.36.3           PFAM.db_3.1.2            
[33] impute_1.42.0             BiocInstaller_1.18.5      GOstats_2.34.0            hwriter_1.3.2            
[37] gtools_3.5.0              BiocParallel_1.2.22       acepack_1.3-3.3           R.oo_1.19.0              
[41] VariantAnnotation_1.14.13 RCurl_1.95-4.7            magrittr_1.5              GO.db_3.1.2              
[45] Formula_1.2-1             oligoClasses_1.30.0       futile.logger_1.4.1       Matrix_1.2-3             
[49] Rcpp_0.12.2               munsell_0.4.2             R.methodsS3_1.7.0         stringi_1.0-1            
[53] zlibbioc_1.14.0           gplots_2.17.0             plyr_1.8.3                grid_3.2.2               
[57] gdata_2.17.0              ReportingTools_2.8.0      lattice_0.20-33           Biostrings_2.36.4        
[61] splines_3.2.2             GenomicFeatures_1.20.6    annotate_1.46.1           locfit_1.5-9.1           
[65] knitr_1.11                geneplotter_1.46.0        reshape2_1.4.1            codetools_0.2-14         
[69] biomaRt_2.24.1            futile.options_1.0.0      XML_3.98-1.3              RcppArmadillo_0.6.400.2.2
[73] biovizBase_1.16.0         latticeExtra_0.6-26       lambda.r_1.1.7            foreach_1.4.3            
[77] gtable_0.1.2              reshape_0.8.5             ggplot2_2.0.0             xtable_1.8-0             
[81] ff_2.2-13                 survival_2.38-3           OrganismDbi_1.10.0        iterators_1.0.8          
[85] GenomicAlignments_1.4.2   AnnotationDbi_1.30.1      cluster_2.0.3             GSEABase_1.30.2     

ADD REPLY •
0
Entering edit mode

Point of clarification: when you say use normalized values in svaseq, are you meaning DESeq2's normalized counts but not edgeR's cpm() function? DESeq2's normalized counts are not integers, but they are unlogged and on the same scale as the original counts, where as normalization via cpm results in counts per million values, which could be logged or unlogged. We use edgeR in our pipelines, so I'm trying to figure out how to integrate svaseq as opposed to regular sva on voom or logCPM values. Would it be best to get unlogged CPM values, then transform to normalized counts using the median library size? Or would using unlogged CPM values be fine?

(I'm helping frodrgs2 here locally, but I've also been using this on other data)

Thanks,

Jenny

ADD REPLY •
2
Entering edit mode
Jeff Leek ▴ 650
@jeff-leek-5015
Last seen 3.7 years ago
United States

Just to be clear, you would have a model matrix consisting of the primary variable you used when estimating the surrogate variables, plus the surrogate variables. Then you would use edgeR to estimate differential expression with respect to the primary variable. 
ADD COMMENT •
0
Entering edit mode
@mrodriguesfernanda-9433
Last seen 7.6 years ago
University of Illinois, Urbana-Champaign

Thank you all for your quick responses.
So just to clarify and check if I am doing this right. I can normalize my data (using edgeR in my case) and then provide normalized cpm data to svaseq, right?

I am including my codes below, not sure if I am doing this right. If someone could give me some insight that would be nice. My data include 24 samples (12 of each genotype - QQ and qq - and each sex - F and M, having 6 biological replicates for each group - QQ.F, QQ.M, qq.F an qq.M). Not sure if I included the estimated surrogate variables in the model properly.

Also, I tried using the normal sva function instead of svaseq and it gave me very similar results. I heard however that the svaseq function does not really account for TMM normalization factors, which would be important for my analysis. Could someone clarify if this is true? Thank you!

My codes:

## QC, FILTERING AND NORMALIZATION USING EDGER ##

> d <- readDGE(targets, labels = targets$Label, comment.char = "#", header = TRUE,
             columns = c(1,7))

> d <- calcNormFactors(d)

> cpm.values <- cpm(d$counts)

> above1cpm <- rowSums(cpm.values  >= 1)

> d.filt <- d[above1cpm >= 6, , keep.lib.sizes = FALSE]

> dim(d.filt)

# [1] 11679      24

> d.filt <- calcNormFactors(d.filt)

> cpm.values.filt < cpm(d.filt$counts)

> head(cpm.values.filt)

(...)
                                           Samples
                 Tags                     QQ.M.3        QQ.M.4       QQ.M.5       QQ.M.6
  ENSBTAG00000020035   99.33248  182.23360  168.80417    94.63499
  ENSBTAG00000011528   48.37818    55.00343    53.30658    62.32472
  ENSBTAG00000012594   17.57829    32.66077    20.34709    18.87298
  ENSBTAG00000018278 242.30761  244.77718  230.29837  241.66867
  ENSBTAG00000021997   13.03218    25.75558    13.58795    19.61574
  ENSBTAG00000008490   17.31310    22.38235    19.68511    23.73471

> logCPM <- cpm(d.filt, log = T)

> head(logCPM)

# (...)
#                                          Samples
#                  Tags                   QQ.M.5     QQ.M.6
#   ENSBTAG00000020035 7.526067 6.557956
#   ENSBTAG00000011528 5.863254 5.955458
#   ENSBTAG00000012594 4.474122 4.232455
#   ENSBTAG00000018278 7.974202 7.910458
#   ENSBTAG00000021997 3.891917 4.288119
#   ENSBTAG00000008490 4.426424 4.562991

## SVA ##

> Group <- relevel(d.filt$samples$group, ref = "qq.M")
> mod <- model.matrix(~Group)
> colnames(mod)[-1] <- paste0(levels(Group)[-1], "vs", levels(Group)[1])
> head(mod)
#   (Intercept) qq.Fvsqq.M  QQ.Fvsqq.M  QQ.Mvsqq.M
# 1              1                  1                0                   0
# 2              1                  1                0                   0
# 3              1                  1                0                   0
# 4              1                  1                0                   0
# 5              1                  1                0                   0
# 6              1                  1                0                   0

> cont.mod <- makeContrasts(QQ.FvsQQ.M = QQ.Fvsqq.M - QQ.Mvsqq.M,
                          qq.Fvsqq.M = qq.Fvsqq.M,
                          QQ.Fvsqq.F = QQ.Fvsqq.M - qq.Fvsqq.M,
                          QQ.Mvsqq.M = QQ.Mvsqq.M,
                          Interact = (QQ.Fvsqq.M - QQ.Mvsqq.M) - (qq.Fvsqq.M),
                          levels = mod)
> cont.mod
# Contrasts
# Levels                 QQ.FvsQQ.M  qq.Fvsqq.M  QQ.Fvsqq.F  QQ.Mvsqq.M  Interact
#         Intercept                       0                  0                  0                    0             0
#     qq.Fvsqq.M                       0                  1                 -1                    0            -1
#    QQ.Fvsqq.M                       1                  0                  1                    0             1
#   QQ.Mvsqq.M                      -1                  0                  0                    1            -1

> svseq = svaseq(cpm.values.filt, mod, mod[,1])
# Number of significant surrogate variables is:  5
# Iteration (out of 5 ):1  2  3  4  5  

> head(svseq$sv)
                    [,1]                [,2]                [,3]                [,4]               [,5]
[1,]  0.05741827 -0.06602954  0.20009240 -0.02684960  0.23766642
[2,]  0.26314274 -0.66218564  0.19546081  0.38882480 -0.21056924
[3,] -0.06735629  0.01038846  0.08790005  0.02410775  0.34376335
[4,]  0.10120300  0.10301322  0.24170332  0.15549638  0.06207581
[5,] -0.17679364 -0.09238517 -0.03882105  0.15047943  0.23952769
[6,]  0.07944103 -0.01603642 -0.26944731 -0.16374470 -0.23344724

> mod.sv <- cbind(mod, svseq$sv)
> colnamesmod.sv)[5:9] <- paste0("sv", 1:5)

#   (Intercept)  qq.Fvsqq. M  QQ.Fvsqq.M  QQ.Mvsqq.M         sv1                      sv2                sv3              sv4              sv5
# 1           1                        1                  0                     0    0.05741827 -0.06602954  0.20009240 -0.02684960  0.23766642
# 2           1                        1                  0                     0    0.26314274 -0.66218564  0.19546081  0.38882480 -0.21056924
# 3           1                        1                  0                     0   -0.06735629  0.01038846  0.08790005  0.02410775  0.34376335
# 4           1                        1                  0                     0    0.10120300  0.10301322  0.24170332  0.15549638  0.06207581
# 5           1                        1                  0                     0   -0.17679364 -0.09238517 -0.03882105  0.15047943  0.23952769
# 6           1                        1                  0                     0    0.07944103 -0.01603642 -0.26944731 -0.16374470 -0.23344724

> cont.mod.sv <- makeContrasts(QQ.FvsQQ.M = QQ.Fvsqq.M - QQ.Mvsqq.M,
                             qq.Fvsqq.M = qq.Fvsqq.M,
                             QQ.Fvsqq.F = QQ.Fvsqq.M - qq.Fvsqq.M,
                             QQ.Mvsqq.M = QQ.Mvsqq.M,
                             Interact = (QQ.Fvsqq.M - QQ.Mvsqq.M) - (qq.Fvsqq.M),
                             levels = mod.sv)

> cont.mod.sv
# Contrasts
# Levels             QQ.FvsQQ.M   qq.Fvsqq.M   QQ.Fvsqq.F   QQ.Mvsqq.M   Interact
#        Intercept                   0                     0                   0                   0              0
#     qq.Fvsqq.M                  0                     1                  -1                   0             -1
#    QQ.Fvsqq.M                  1                     0                   1                   0              1
#   QQ.Mvsqq.M                 -1                     0                   0                   1             -1
#                  sv1                 0                      0                   0                   0              0
#                  sv2                 0                      0                   0                   0              0
#                  sv3                 0                      0                   0                   0              0
#                  sv4                 0                      0                   0                   0              0
#                  sv5                 0                      0                   0                   0              0

> d.filt.sv <- estimateGLMCommonDisp(d.filt, mod.sv, verbose = TRUE)
# Disp = 0.0268 , BCV = 0.1637    ###not sure if the BCV value is good. I am working on bovine. edgeR recommends
                                                              0.1 for identical strains and 0.4 for humans. Not sure if 0.1637 is okay for bovine.

> d.filt.sv <- estimateGLMTrendedDisp(d.filt, mod.sv)
> d.filt.sv <- estimateGLMTagwiseDisp(d.filt, mod.sv)

> plotBCVd.filt.sv)

> fit.edgeR.sv <- glmFitd.filt.sv, mod.sv)
 

> eR.QQ.FvsQQ.M.sv <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 1])
> eR.qq.Fvsqq.M.sv   <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 2])
> eR.QQ.Fvsqq.F.sv   <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 3])
> eR.QQ.Mvsqq.M.sv <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 4])
> eR.Interact.sv <- glmLRTfit.edgeR.sv, contrast = cont.mod.sv[ , 5])

# I am not sure on how to check on sv significance from here or if what I did above is right. Should I have the svs in the model another way? Any insight are welcome! Thank you again

R version 3.2.2 (2015-08-14)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu precise (12.04.5 LTS)

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats4    parallel  stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] sva_3.14.0           genefilter_1.50.0    mgcv_1.8-10          nlme_3.1-122         rtracklayer_1.28.10
 [6] GenomicRanges_1.20.8 GenomeInfoDb_1.4.3   IRanges_2.2.9        S4Vectors_0.6.6      rgl_0.95.1429       
[11] edgeR_3.10.5         limma_3.24.15        affycoretools_1.40.5 affy_1.46.1          Biobase_2.28.0      
[16] BiocGenerics_0.14.0  WGCNA_1.48           RSQLite_1.0.0        DBI_0.3.1            fastcluster_1.1.16  
[21] dynamicTreeCut_1.62

loaded via a namespace (and not attached):
 [1] Category_2.34.2           bitops_1.0-6              matrixStats_0.50.1        doParallel_1.0.10        
 [5] RColorBrewer_1.1-2        gcrma_2.40.0              tools_3.2.2               affyio_1.36.0            
 [9] KernSmooth_2.23-15        rpart_4.1-10              Hmisc_3.17-1              colorspace_1.2-6         
[13] nnet_7.3-11               gridExtra_2.0.0           GGally_0.5.0              DESeq2_1.8.2             
[17] bit_1.1-12                preprocessCore_1.30.0     graph_1.46.0              ggbio_1.16.1             
[21] caTools_1.17.1            scales_0.3.0              RBGL_1.44.0               stringr_1.0.0            
[25] Rsamtools_1.20.5          foreign_0.8-66            R.utils_2.2.0             AnnotationForge_1.10.1   
[29] XVector_0.8.0             dichromat_2.0-0           BSgenome_1.36.3           PFAM.db_3.1.2            
[33] impute_1.42.0             BiocInstaller_1.18.5      GOstats_2.34.0            hwriter_1.3.2            
[37] gtools_3.5.0              BiocParallel_1.2.22       acepack_1.3-3.3           R.oo_1.19.0              
[41] VariantAnnotation_1.14.13 RCurl_1.95-4.7            magrittr_1.5              GO.db_3.1.2              
[45] Formula_1.2-1             oligoClasses_1.30.0       futile.logger_1.4.1       Matrix_1.2-3             
[49] Rcpp_0.12.2               munsell_0.4.2             R.methodsS3_1.7.0         stringi_1.0-1            
[53] zlibbioc_1.14.0           gplots_2.17.0             plyr_1.8.3                grid_3.2.2               
[57] gdata_2.17.0              ReportingTools_2.8.0      lattice_0.20-33           Biostrings_2.36.4        
[61] splines_3.2.2             GenomicFeatures_1.20.6    annotate_1.46.1           locfit_1.5-9.1           
[65] knitr_1.11                geneplotter_1.46.0        reshape2_1.4.1            codetools_0.2-14         
[69] biomaRt_2.24.1            futile.options_1.0.0      XML_3.98-1.3              RcppArmadillo_0.6.400.2.2
[73] biovizBase_1.16.0         latticeExtra_0.6-26       lambda.r_1.1.7            foreach_1.4.3            
[77] gtable_0.1.2              reshape_0.8.5             ggplot2_2.0.0             xtable_1.8-0             
[81] ff_2.2-13                 survival_2.38-3           OrganismDbi_1.10.0        iterators_1.0.8          
[85] GenomicAlignments_1.4.2   AnnotationDbi_1.30.1      cluster_2.0.3             GSEABase_1.30.2     

 
ADD COMMENT •

Login before adding your answer.

Similar Posts
Loading Similar Posts
Traffic: 838 users visited in the last hour
Content Search
Users
Tags
Badges
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6