DESeq2 relevel question and resultsNames
Entering edit mode
Hannah • 0
Last seen 24 days ago
United States

I am having trouble understanding why my resultsNames are missing comparisons that I thought should be there, as well as if I should relevel for both factors in my analysis?

I have two treatments 1-Year of sample and 2-Timpoint of collection. I am conducting an LRT test.

What I want to know is, shouldn't there be a "Year_H2019_vs_H1999", "Timepoint_T24_vs_T48" , "YearH1999.TimepointT0", "YearH2019.TimepointT0", "YearH1995.TimepointT0", "YearH1995.TimepointT24", and a "YearH1995.TimepointT48", present in the results(Names)?

My second question is: I want to have comparisons based on timepoint 0, but do I also need to relevel on the "Year" factor?

Thanks for any guidance on these problems.

#Here is my metadata
      Year Timepoint    Group
gm10 H1999       T24 1999_T24
gm11 H1995       T24 1995_T24
gm12 H1995       T24 1995_T24
gm13 H1999       T24 1999_T24
gm14 H1999       T48 1999_T48
gm15 H2019       T48 2019_T48
gm16 H1999       T48 1999_T48
gm17 H2019       T24 2019_T24
gm18 H1995       T24 1995_T24
gm19 H1999        T0  1999_T0
gm1  H2019       T24 2019_T24
gm20 H1995        T0  1995_T0
gm21 H1999        T0  1999_T0
gm22 H1999        T0  1999_T0
gm23 H1995        T0  1995_T0
gm24 H1999       T48 1999_T48
gm25 H2019       T48 2019_T48
gm26 H1999       T48 1999_T48
gm27 H1995       T48 1995_T48
gm28 H2019       T24 2019_T24
gm29 H1999        T0  1999_T0
gm2  H2019        T0  2019_T0
gm30 H1999       T24 1999_T24
gm31 H1995        T0  1995_T0
gm32 H2019        T0  2019_T0
gm33 H1999       T24 1999_T24
gm34 H2019       T48 2019_T48
gm35 H1995       T24 1995_T24
gm36 H2019       T24 2019_T24
gm3  H1995        T0  1995_T0
gm4  H1995       T48 1995_T48
gm5  H2019       T48 2019_T48
gm6  H1995       T48 1995_T48
gm7  H2019        T0  2019_T0
gm8  H1995       T48 1995_T48
gm9  H2019        T0  2019_T0

##My model 
dds<- DESeqDataSetFromMatrix(countData = mycounts, colData = metadata, design=~Year+Timepoint+Year:Timepoint)
##relevel for Timepoint 0 to be the reference
dds$Timepoint<- relevel(dds$Timepoint, ref= "T0")

dds<- DESeq(dds)
res <- results(dds)
> resultsNames(dds)
[1] "Intercept"              "Year_H1999_vs_H1995"    "Year_H2019_vs_H1995"    "Timepoint_T24_vs_T0"   
[5] "Timepoint_T48_vs_T0"    "YearH1999.TimepointT24" "YearH2019.TimepointT24" "YearH1999.TimepointT48"
[9] "YearH2019.TimepointT48"

# include your problematic code here with any corresponding output 
# please also include the results of running the following in an R session 

> sessionInfo( )
R version 4.0.1 (2020-06-06)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows >= 8 x64 (build 9200)

Matrix products: default

[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

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

other attached packages:
 [1] forcats_0.5.1               stringr_1.4.0               purrr_0.3.4                
 [4] readr_2.1.2                 tidyr_1.2.0                 tibble_3.1.6               
 [7] ggplot2_3.3.5               tidyverse_1.3.1             DESeq2_1.28.1              
[10] SummarizedExperiment_1.18.2 DelayedArray_0.14.1         matrixStats_0.61.0         
[13] Biobase_2.48.0              GenomicRanges_1.40.0        GenomeInfoDb_1.24.2        
[16] IRanges_2.22.2              S4Vectors_0.26.1            BiocGenerics_0.34.0        
[19] dplyr_1.0.8                

loaded via a namespace (and not attached):
 [1] httr_1.4.2             bit64_4.0.5            jsonlite_1.8.0         splines_4.0.1         
 [5] modelr_0.1.8           assertthat_0.2.1       blob_1.2.2             cellranger_1.1.0      
 [9] GenomeInfoDbData_1.2.3 yaml_2.2.2             pillar_1.7.0           RSQLite_2.2.10        
[13] backports_1.4.1        lattice_0.20-45        glue_1.6.1             RColorBrewer_1.1-3    
[17] XVector_0.28.0         rvest_1.0.2            colorspace_2.0-2       Matrix_1.4-0          
[21] XML_3.99-0.8           pkgconfig_2.0.3        broom_0.7.12           haven_2.4.3           
[25] genefilter_1.70.0      zlibbioc_1.34.0        xtable_1.8-4           scales_1.1.1          
[29] tzdb_0.2.0             BiocParallel_1.22.0    annotate_1.66.0        generics_0.1.2        
[33] ellipsis_0.3.2         withr_2.5.0            cachem_1.0.6           cli_3.1.1             
[37] survival_3.2-13        magrittr_2.0.2         crayon_1.5.1           readxl_1.3.1          
[41] memoise_2.0.1          fs_1.5.2               fansi_1.0.2            xml2_1.3.3            
[45] tools_4.0.1            hms_1.1.1              lifecycle_1.0.1        reprex_2.0.1          
[49] munsell_0.5.0          locfit_1.5-9.4         AnnotationDbi_1.50.3   compiler_4.0.1        
[53] rlang_1.0.1            grid_4.0.1             RCurl_1.98-1.6         rstudioapi_0.13       
[57] bitops_1.0-7           gtable_0.3.0           DBI_1.1.2              R6_2.5.1              
[61] lubridate_1.8.0        fastmap_1.1.0          bit_4.0.4              utf8_1.2.2            
[65] stringi_1.7.6          Rcpp_1.0.8             vctrs_0.3.8            geneplotter_1.66.0    
[69] dbplyr_2.1.1           tidyselect_1.1.2
DESeq2 • 181 views
Entering edit mode
Last seen 6 hours ago
United States

I am having trouble understanding why my resultsNames are missing comparisons that I thought should be there, as well as if I should relevel for both factors in my analysis?

For questions about statistical design and interpretation of results, I recommend consulting a local statistician or someone familiar with linear models in R. I have to reserve my time on the support site for software related questions.

Entering edit mode

I apologize, I thought this was a software question on releveling and missing comparisons. Maybe someone else could share a link that has a good example or documentation of a two factor analysis that includes releveling? I have been trying to find thorough documentation, but haven't been able to find a good example.

Entering edit mode

In R, one can specify a particular linear model using formula and this produces a set of coefficients: this is generic to all methods in Bioconductor and beyond. So anyone who knows how to construct and interpret the design and coefficients in R would be able to help you. I try to therefore reserve my time on the support site to specific questions about DESeq2 software. There are a lot of users online with general statistical consulting questions about how to analyze their datasets and interpret results, but I just don't have time to address all of those questions about experimental design here, while also maintaining and supporting software.


Login before adding your answer.

Traffic: 280 users visited in the last hour
Help About
Access RSS

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

Powered by the version 2.3.6