Log fold change interpretation of time series with Wald test
Entering edit mode
jd348 • 0
Last seen 1 day ago


I have a question regarding the results output of the Wald test I'm running in DESeq2. I have time-series data with samples collected at 5 time points b/w 2 genotypes: 0hr, 2hr, 6hr, 12hr, 24hr and mutant vs. control. My question is this: using the design in my code below, how is the Log2FoldChange calculated in the results table, considering that there are multiple time points with every gene? Is it the average log2 fold change across all time points per gene?

dds <- DESeqDataSetFromMatrix(countData = ctdata, colData = cldata, design = ~ genotype + time)
dds <- dds[rowSums(counts(dds) > 10,]
dds_res <- results(dds)

sessionInfo( )
R version 4.1.0 (2021-05-18)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Big Sur 11.4

Matrix products: default
LAPACK: /Library/Frameworks/R.framework/Versions/4.1/Resources/lib/libRlapack.dylib

[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

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

other attached packages:
 [1] DEGreport_1.28.0            gplots_3.1.1                VennDiagram_1.6.20          futile.logger_1.4.3         WGCNA_1.70-3               
 [6] fastcluster_1.2.3           dynamicTreeCut_1.63-1       WebGestaltR_0.4.4           pheatmap_1.0.12             ggpubr_0.4.0               
[11] openxlsx_4.2.4              BiocParallel_1.26.2         ggrepel_0.9.1               forcats_0.5.1               stringr_1.4.0              
[16] dplyr_1.0.7                 purrr_0.3.4                 readr_2.0.1                 tidyr_1.1.3                 tibble_3.1.4               
[21] ggplot2_3.3.5               tidyverse_1.3.1             DESeq2_1.32.0               SummarizedExperiment_1.22.0 Biobase_2.52.0             
[26] MatrixGenerics_1.4.3        matrixStats_0.60.1          GenomicRanges_1.44.0        GenomeInfoDb_1.28.2         IRanges_2.26.0             
[31] S4Vectors_0.30.0            BiocGenerics_0.38.0        

loaded via a namespace (and not attached):
  [1] utf8_1.2.2                  tidyselect_1.1.1            RSQLite_2.2.8               AnnotationDbi_1.54.1        htmlwidgets_1.5.3          
  [6] munsell_0.5.0               codetools_0.2-18            preprocessCore_1.54.0       withr_2.4.2                 colorspace_2.0-2           
 [11] knitr_1.33                  rstudioapi_0.13             ggsignif_0.6.2              labeling_0.4.2              lasso2_1.2-21.1            
 [16] GenomeInfoDbData_1.2.6      mnormt_2.0.2                farver_2.1.0                bit64_4.0.5                 vctrs_0.3.8                
 [21] generics_0.1.0              lambda.r_1.2.4              xfun_0.25                   R6_2.5.1                    doParallel_1.0.16          
 [26] clue_0.3-59                 locfit_1.5-9.4              reshape_0.8.8               bitops_1.0-7                cachem_1.0.6               
 [31] DelayedArray_0.18.0         assertthat_0.2.1            vroom_1.5.4                 scales_1.1.1                nnet_7.3-16                
 [36] gtable_0.3.0                Cairo_1.5-12.2              rlang_0.4.11                genefilter_1.74.0           systemfonts_1.0.2          
 [41] GlobalOptions_0.1.2         rstatix_0.7.0               impute_1.66.0               broom_0.7.9                 checkmate_2.0.0            
 [46] abind_1.4-5                 modelr_0.1.8                backports_1.2.1             Hmisc_4.5-0                 tools_4.1.0                
 [51] psych_2.1.6                 logging_0.10-108            ellipsis_0.3.2              RColorBrewer_1.1-2          ggdendro_0.1.22            
 [56] plyr_1.8.6                  Rcpp_1.0.7                  base64enc_0.1-3             zlibbioc_1.38.0             RCurl_1.98-1.4             
 [61] rpart_4.1-15                GetoptLong_1.0.5            cowplot_1.1.1               haven_2.4.3                 cluster_2.1.2              
 [66] fs_1.5.0                    apcluster_1.4.8             magrittr_2.0.1              data.table_1.14.0           futile.options_1.0.1       
 [71] circlize_0.4.13             reprex_2.0.1                tmvnsim_1.0-2               whisker_0.4                 hms_1.1.0                  
 [76] xtable_1.8-4                XML_3.99-0.7                rio_0.5.27                  jpeg_0.1-9                  readxl_1.3.1               
 [81] gridExtra_2.3               shape_1.4.6                 compiler_4.1.0              KernSmooth_2.23-20          crayon_1.4.1               
 [86] htmltools_0.5.2             mgcv_1.8-36                 tzdb_0.1.2                  Formula_1.2-4               geneplotter_1.70.0         
 [91] lubridate_1.7.10            DBI_1.1.1                   formatR_1.11                dbplyr_2.1.1                ComplexHeatmap_2.8.0       
 [96] MASS_7.3-54                 Matrix_1.3-4                car_3.0-11                  cli_3.0.1                   igraph_1.2.6               
[101] pkgconfig_2.0.3             foreign_0.8-81              xml2_1.3.2                  foreach_1.5.1               svglite_2.0.0              
[106] annotate_1.70.0             rngtools_1.5                XVector_0.32.0              rvest_1.0.1                 doRNG_1.8.2                
[111] digest_0.6.27               ConsensusClusterPlus_1.56.0 Biostrings_2.60.2           cellranger_1.1.0            htmlTable_2.2.1            
[116] edgeR_3.34.0                curl_4.3.2                  gtools_3.9.2                rjson_0.2.20                nlme_3.1-152               
[121] lifecycle_1.0.0             jsonlite_1.7.2              carData_3.0-4               limma_3.48.3                fansi_0.5.0                
[126] pillar_1.6.2                lattice_0.20-44             Nozzle.R1_1.1-1             KEGGREST_1.32.0             fastmap_1.1.0              
[131] httr_1.4.2                  survival_3.2-13             GO.db_3.13.0                glue_1.4.2                  zip_2.2.0                  
[136] png_0.1-7                   iterators_1.0.13            bit_4.0.4                   stringi_1.7.4               blob_1.2.2                 
[141] latticeExtra_0.6-29         caTools_1.18.2              memoise_2.0.0
DESeq2 • 187 views
Entering edit mode
swbarnes2 ▴ 890
Last seen 53 minutes ago
San Diego

In plain, not statistically technical language:

When you run DESeq2 with a series of numerical colData like time or dosage, what's returned is not a Log2fold change, but the slope of the line of the normalized, logged data.

Entering edit mode

So would you recommend still using Log2fold change as a measurement for effect size?


Login before adding your answer.

Traffic: 308 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