Search
Question: Difference between DESeq2 1.14 and 1.16
0
gravatar for maksims.fiosins
6 days ago by
maksims.fiosins0 wrote:

Hello,

I have two versions of DESeq2, 1.14 and 1.16. I run the following code:

dds <- makeExampleDESeqDataSet(n=100,m=12)
dds$genotype <- factor(rep(rep(c("I","II"),each=3),2))
design(dds) <- ~ genotype + condition
dds <- DESeq(dds)
resultsNames(dds)

For the version 1.14 it produces the following output:

[1] "Intercept"  "genotypeI"  "genotypeII" "conditionA" "conditionB"

And for the version 1.16 the following:

[1] "Intercept"        "genotype_II_vs_I" "condition_B_vs_A"

My question is, if on the version 1.16 it is possible to revert to the 1.14 output (single levels of factors) in order to make analysis like DESeq2 likelihood ratio test (LRT) design - 2 genotypes, 4 time points?

Session info of the 1.14:

R version 3.3.1 (2016-06-21)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 14.04.5 LTS

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

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

other attached packages:
[1] DESeq2_1.14.1              SummarizedExperiment_1.4.0
[3] Biobase_2.34.0             GenomicRanges_1.26.4      
[5] GenomeInfoDb_1.10.3        IRanges_2.8.2             
[7] S4Vectors_0.12.2           BiocGenerics_0.20.0       

loaded via a namespace (and not attached):
 [1] genefilter_1.56.0    locfit_1.5-9.1       splines_3.3.1       
 [4] lattice_0.20-35      colorspace_1.3-2     htmltools_0.3.6     
 [7] base64enc_0.1-3      blob_1.1.0           survival_2.41-3     
[10] XML_3.98-1.9         rlang_0.1.1          foreign_0.8-69      
[13] DBI_0.7              BiocParallel_1.8.2   bit64_0.9-7         
[16] RColorBrewer_1.1-2   plyr_1.8.4           stringr_1.2.0       
[19] zlibbioc_1.20.0      munsell_0.4.3        gtable_0.2.0        
[22] htmlwidgets_0.8      memoise_1.1.0        latticeExtra_0.6-28
[25] knitr_1.16           geneplotter_1.52.0   AnnotationDbi_1.36.2
[28] htmlTable_1.9        Rcpp_0.12.11         acepack_1.4.1       
[31] xtable_1.8-2         scales_0.4.1         backports_1.1.0     
[34] checkmate_1.8.3      Hmisc_4.0-3          annotate_1.52.1     
[37] XVector_0.14.1       bit_1.1-12           gridExtra_2.2.1     
[40] ggplot2_2.2.1        digest_0.6.12        stringi_1.1.5       
[43] grid_3.3.1           bitops_1.0-6         tools_3.3.1         
[46] magrittr_1.5         lazyeval_0.2.0       RCurl_1.95-4.8      
[49] tibble_1.3.3         RSQLite_2.0          Formula_1.2-1       
[52] cluster_2.0.6        Matrix_1.2-10        data.table_1.10.4   
[55] rpart_4.1-11         nnet_7.3-12         

Session info of the 1.16:

R version 3.4.0 (2017-04-21)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Red Hat Enterprise Linux

Matrix products: default
BLAS: /usr/local/lib64/R/lib/libRblas.so
LAPACK: /usr/local/lib64/R/lib/libRlapack.so

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

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

other attached packages:
 [1] DESeq2_1.16.1              SummarizedExperiment_1.6.1
 [3] DelayedArray_0.2.2         matrixStats_0.52.2        
 [5] Biobase_2.36.2             GenomicRanges_1.28.2      
 [7] GenomeInfoDb_1.12.0        IRanges_2.10.1            
 [9] S4Vectors_0.14.1           BiocGenerics_0.22.0       

loaded via a namespace (and not attached):
 [1] genefilter_1.58.1       locfit_1.5-9.1          splines_3.4.0          
 [4] lattice_0.20-35         colorspace_1.3-2        htmltools_0.3.6        
 [7] base64enc_0.1-3         survival_2.41-3         XML_3.98-1.7           
[10] rlang_0.1.1             DBI_0.6-1               foreign_0.8-68         
[13] BiocParallel_1.10.1     RColorBrewer_1.1-2      GenomeInfoDbData_0.99.0
[16] plyr_1.8.4              stringr_1.2.0           zlibbioc_1.22.0        
[19] munsell_0.4.3           gtable_0.2.0            htmlwidgets_0.8        
[22] memoise_1.1.0           latticeExtra_0.6-28     knitr_1.15.1           
[25] geneplotter_1.54.0      AnnotationDbi_1.38.0    htmlTable_1.9          
[28] Rcpp_0.12.10            acepack_1.4.1           xtable_1.8-2           
[31] scales_0.4.1            backports_1.0.5         checkmate_1.8.2        
[34] Hmisc_4.0-3             annotate_1.54.0         XVector_0.16.0         
[37] gridExtra_2.2.1         ggplot2_2.2.1           digest_0.6.12          
[40] stringi_1.1.5           grid_3.4.0              tools_3.4.0            
[43] bitops_1.0-6            magrittr_1.5            RSQLite_1.1-2          
[46] lazyeval_0.2.0          RCurl_1.95-4.8          tibble_1.3.1           
[49] Formula_1.2-1           cluster_2.0.6           Matrix_1.2-10          
[52] data.table_1.10.4       rpart_4.1-11            nnet_7.3-12            
[55] compiler_3.4.0    
ADD COMMENTlink modified 6 days ago by Gavin Kelly250 • written 6 days ago by maksims.fiosins0
2
gravatar for Gavin Kelly
6 days ago by
Gavin Kelly250
United Kingdom / London / Francis Crick Institute
Gavin Kelly250 wrote:
You can use
dds <- DESeq(dds, betaPrior=TRUE)

to get the resultsNames to match, but I think if you're doing an LRT test, then it shouldn't make much difference, as the full and reduced models are specified in exactly the same way as before, without referencing specific coefficients.  And if you're doing a Wald test, the same portfolio of contrasts is available (though you may have to 'reword' them in terms of the new, relative, resultsNames), and you can moderate the fold-change (as was done by default in earlier versions of DESeq2) with the lfcShrink function.  I think an advantage of the new approach is that one can shrink the specific lfc one is interested in, which is a good thing.

 

ADD COMMENTlink written 6 days ago by Gavin Kelly250
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 227 users visited in the last hour