I am trying to use DRIMseq for DTU with 2 treatments on two different strains of mice. In my previous analysis with genes only, I had found a strong interaction effect of strain and treatments. Is it possible to model that with DRIMseq?

Also when I try to add strain term and use it as the batch effect as described in the tutorial, I get the following error:

! Using a subset of 0.1 genes to estimate common precision !
Error in optimHess(par = par, fn = dm_lik_regG, gr = dm_score_regG, x = x,  : 
  non-finite value supplied by optim
In addition: There were 50 or more warnings (use warnings() to see the first 50)

Please see the sessionInfo() below.

R version 3.4.0 (2017-04-21)
Platform: x86_64-redhat-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)

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

 [1] LC_CTYPE=en_GB.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_GB.UTF-8     LC_MONETARY=en_US.UTF-8   
 [6] LC_MESSAGES=en_GB.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                  LC_ADDRESS=C               LC_TELEPHONE=C            

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

other attached packages:
 [1] DRIMSeq_1.6.0                            pheatmap_1.0.8                           ggplot2_2.2.1                           
 [4] edgeR_3.16.5                             limma_3.30.13                            TxDb.Mmusculus.UCSC.mm10.knownGene_3.4.0
 [7] GenomicFeatures_1.26.4                   AnnotationDbi_1.36.2                     Biobase_2.34.0                          
[10] GenomicRanges_1.26.4                     GenomeInfoDb_1.10.3                      IRanges_2.8.2                           
[13] S4Vectors_0.12.2                         BiocGenerics_0.20.0                      tximport_1.2.0                          

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.12               RColorBrewer_1.1-2         plyr_1.8.4                 pillar_1.2.1               compiler_3.4.0            
 [6] XVector_0.14.1             bitops_1.0-6               tools_3.4.0                zlibbioc_1.20.0            biomaRt_2.30.0            
[11] digest_0.6.12              bit_1.1-12                 gtable_0.2.0               RSQLite_2.0                memoise_1.1.0             
[16] tibble_1.4.2               lattice_0.20-35            pkgconfig_2.0.1            rlang_0.2.0                Matrix_1.2-10             
[21] DBI_0.7                    stringr_1.3.0              rtracklayer_1.34.2         Biostrings_2.42.1          locfit_1.5-9.1            
[26] bit64_0.9-7                grid_3.4.0                 XML_3.98-1.10              BiocParallel_1.8.2         magrittr_1.5              
[31] reshape2_1.4.3             blob_1.1.0                 MASS_7.3-47                scales_0.5.0               Rsamtools_1.26.2          
[36] GenomicAlignments_1.10.1   SummarizedExperiment_1.4.0 colorspace_1.3-2           labeling_0.3               stringi_1.1.7             
[41] lazyeval_0.2.1             munsell_0.4.3              RCurl_1.95-4.10           

One way to see the interaction effect would be including it in the design matrix as explained in the limma tutorial. Something on the line of:

design_full <- model.matrix(~ treatment*strain, data = samples(d))

Or you could build it yourself in a matrix like

 (Intercept) strainA_treatmentA strainA_treatmentB  strainB_treatmentA strainB_treatmentB
       1                      0                  1                  0                   0
       1                      1                  0                  0                   0

About the error, it seems like the data is not in the right format, you have either NA or Inf somewhere in the matrix. Check it and rethink the processing you made.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.