Multiple groups comparison using Linnorm
Entering edit mode
awakumaya • 0
Last seen 2.7 years ago

Hi all,

I'm using Linnorm for multiple groups comparison. I succeeded 2 groups comparison, such as WT vs KO, by following example commands that Linnorm developers showed. The commands were the below:

#Obtain example matrix:
#Create limma design matrix (first 5 columns are tumor, last 5 columns are normal)
designmatrix <- c(rep(1,5),rep(2,5))
designmatrix <- model.matrix(~ 0+factor(designmatrix))
colnames(designmatrix) <- c("group1", "group2")
rownames(designmatrix) <- colnames(LIHC)
#DEG analysis
DEGResults <- Linnorm.limma(LIHC, designmatrix)

And I got a same output to the one that the developer mentioned in Package 'Linnorm' documents. The results is the below:

                            logFC          XPM          t      P.Value    adj.P.Val             B
CACNA1S|779             0.6931234 6.513083e-03  43.944443 8.127946e-12 1.580235e-07 11.7115287634
HOXD4|3233              2.9022970 6.838703e-02  40.271851 1.776754e-11 1.727182e-07 11.5495184015
PYY|5697                2.0850513 3.424445e-02  37.493428 3.369902e-11 2.183921e-07 11.3990864805
: : : :

So, I tried anova analysis of 3 groups using the same data by changing design matrix, like this.

designmatrix <- c(rep(1,3),rep(2,4), rep(3,3))
designmatrix <- model.matrix(~ 0+factor(designmatrix))
colnames(designmatrix) <- c("group1", "group2", "group3")
rownames(designmatrix) <- colnames(LIHC)

and then do "Linnorm.limma(LIHC, designmatrix)", like at 2-groups comparion. But the output has some different columns like this.

                              XPM group2.group3   AveExpr          F      P.Value    adj.P.Val
PSMB11|122706        1.148511e-02   0.000000000  2.153807 1847.32855 2.117622e-11 3.365242e-07
A2ML1|144568         3.700802e-02   0.000000000  2.474972 1633.54576 3.461827e-11 3.365242e-07
FMO9P|116123         3.481089e-02   0.000000000  2.449994 1323.63062 8.021902e-11 4.019964e-07
: : : :

Were a serial of my commands at the 3-groups analysis correct and the P.value and adj.P.Val in the results are reliable? If it's so, what do the value in the second column mean? Why group2 and group3 were picked up in the all?

> sessionInfo( )
R version 4.1.0 (2021-05-18)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19041)

Matrix products: default

[1] LC_COLLATE=Japanese_Japan.932  LC_CTYPE=Japanese_Japan.932    LC_MONETARY=Japanese_Japan.932
[4] LC_NUMERIC=C                   LC_TIME=Japanese_Japan.932    

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

other attached packages:
[1] Linnorm_2.16.0    limma_3.48.1      reshape_0.8.8     data.table_1.14.0

loaded via a namespace (and not attached):
 [1] fastcluster_1.2.3        gtools_3.9.2             statmod_1.4.36          
 [4] zoo_1.8-9                modeltools_0.2-23        tidyselect_1.1.1        
 [7] kernlab_0.9-29           purrr_0.3.4              splines_4.1.0           
[10] lattice_0.20-44          colorspace_2.0-2         vctrs_0.3.8             
[13] generics_0.1.0           amap_0.8-18              stats4_4.1.0            
[16] mgcv_1.8-36              utf8_1.2.2               rlang_0.4.11            
[19] pillar_1.6.1             glue_1.4.2               DBI_1.1.1               
[22] prabclus_2.3-2           fpc_2.2-9                lifecycle_1.0.0         
[25] plyr_1.8.6               robustbase_0.93-8        munsell_0.5.0           
[28] gtable_0.3.0             RcppArmadillo_0. permute_0.9-5           
[31] flexmix_2.3-17           curl_4.3.2               class_7.3-19            
[34] fansi_0.5.0              DEoptimR_1.0-9           Rcpp_1.0.7              
[37] diptest_0.76-0           scales_1.1.1             BiocManager_1.30.16     
[40] gdata_2.18.0             vegan_2.5-7              apcluster_1.4.8         
[43] ellipse_0.4.2            ggplot2_3.3.5            gmodels_2.18.1          
[46] Rtsne_0.15               dplyr_1.0.7              grid_4.1.0              
[49] tools_4.1.0              magrittr_2.0.1           tibble_3.1.3            
[52] cluster_2.1.2            ggdendro_0.1.22          crayon_1.4.1            
[55] pkgconfig_2.0.3          Matrix_1.3-4             ellipsis_0.3.2          
[58] MASS_7.3-54              assertthat_0.2.1         R6_2.5.0                
[61] mclust_5.4.7             nlme_3.1-152             nnet_7.3-16             
[64] igraph_1.2.6             compiler_4.1.0
Linnorm DEG MultipleComparison • 674 views
Entering edit mode
shunyip • 0
Last seen 6 months ago
United States

Sorry for the late reply. I didn't notice a post here.

This matrix is an edited limma result output. So, the p values are still correct, but the first few columns are wrongly edited by Linnorm. It looks like limma's output has changed since I first incorporated limma into Linnorm.

One alternative is to feed Linnorm's transformed data directly into limma, which is what this function does.

I have fixed this issue at version 2.25.5. Thanks for bringing this up.


Login before adding your answer.

Traffic: 445 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