Search
Question: No gene hits in cameraPR - due to bug dropping statistic names?
1
gravatar for sarah.williams1
16 days ago by
sarah.williams120 wrote:

Hi,

I have been using the pre-ranked cameraPR function for a custom ranking of genes for camera. 

If I use the function as is, I get no hit genes at all:

cameraPR(statistic=the_stat,index=Hs.c5.gene_name)[1:3,]
                                              NGenes Direction PValue FDR
GO_REGULATION_OF_DOPAMINE_METABOLIC_PROCESS        0        Up    NaN NaN
GO_LACTATE_TRANSPORT                               0        Up    NaN NaN
GO_POSITIVE_REGULATION_OF_VIRAL_TRANSCRIPTION      0        Up    NaN NaN

 

I think this might be due to gene names of the statistic getting stripped away inside the function by as.numeric(). Simple fix - If I make the following change to the cameraPR.default it appears to work:

#    Check statistic
if(anyNA(statistic)) stop("NA values for statistic now allowed")
Stat <- as.numeric(statistic)
G <- length(Stat)
# ID <- names(Stat)    # Replace Stat with statistic which still has names
ID <- names(statistic) 
if(G<3) stop("Two few genes in dataset: need at least 3")

 

Output after change:

cameraPR.default.edit(statistic=the_stat,index=Hs.c5.gene_name)[1:3,]
                               NGenes Direction       PValue          FDR
GO_STEROL_BIOSYNTHETIC_PROCESS     32        Up 2.850148e-18 1.754266e-14
GO_MICROGLIAL_CELL_ACTIVATION       4        Up 3.279112e-13 1.009147e-09
GO_CYTOSOLIC_RIBOSOME              96        Up 8.743939e-12 1.793965e-08

So I think it is a bug, but if it is a case of user error please let me know! I've just redefined a function locally as a workaround. 

Also, please advise if there is somewhere else I should be reporting this - primary repo or bug tracker is ? (SVN? github? here?)

Thanks,

Sarah.

 

-----------------------------------------------------------------------------------------------------------

# Input format is like:

#(NB: Ignore the 0s, I'm still figuring out the statistic)

the_stat[1:10]
   IGFBP3      RERG     STMN4    IGSF21 LINC00890     CDH11      GNG3    INSIG1       JUN      SGK1 
 0.000000  0.000000  9.443428  9.031877  0.000000  0.000000  8.835153  8.420033  8.168065  0.000000 
head(Hs.c5.gene_name)
$GO_REGULATION_OF_DOPAMINE_METABOLIC_PROCESS
 [1] "ABAT"   "CHRNB2" "COMT"   "DRD1"   "DRD4"   "GPR37"  "HPRT1"  "HTR1A"  "MAOB"   "NR4A2"  "PARK2"  "PARK7" 
[13] "PDE1B"  "PNKD"   "SLC6A3" "SNCA"   "TACR3" 
$GO_LACTATE_TRANSPORT
 [1] "EMB"      "SLC16A1"  "SLC16A11" "SLC16A12" "SLC16A13" "SLC16A3"  "SLC16A4"  "SLC16A5"  "SLC16A6"  "SLC16A7" 
[11] "SLC16A8"  "SLC5A12" 

 

> sessionInfo()
R version 3.4.1 (2017-06-30)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 16.04.2 LTS

Matrix products: default
BLAS: /mnt/software/apps/R/3.4.1/lib/R/lib/libRblas.so
LAPACK: /mnt/software/apps/R/3.4.1/lib/R/lib/libRlapack.so

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

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

other attached packages:
 [1] gplots_3.0.1              reshape2_1.4.2            EnsDb.Hsapiens.v79_2.1.0  BiocInstaller_1.26.1     
 [5] EnsDb.Mmusculus.v79_2.1.0 ensembldb_2.0.4           AnnotationFilter_1.0.0    GenomicFeatures_1.28.4   
 [9] AnnotationDbi_1.38.1      Biobase_2.36.2            GenomicRanges_1.28.3      GenomeInfoDb_1.12.2      
[13] IRanges_2.10.2            S4Vectors_0.14.3          BiocGenerics_0.22.0       edgeR_3.18.1             
[17] limma_3.32.5             

loaded via a namespace (and not attached):
 [1] SummarizedExperiment_1.6.3    gtools_3.5.0                  locfit_1.5-9.1               
 [4] lattice_0.20-35               htmltools_0.3.6               rtracklayer_1.36.4           
 [7] yaml_2.1.14                   interactiveDisplayBase_1.14.0 blob_1.1.0                   
[10] XML_3.98-1.9                  rlang_0.1.1                   DBI_0.7                      
[13] BiocParallel_1.10.1           bit64_0.9-7                   plyr_1.8.4                   
[16] matrixStats_0.52.2            GenomeInfoDbData_0.99.0       stringr_1.2.0                
[19] zlibbioc_1.22.0               ProtGenerics_1.8.0            Biostrings_2.44.2            
[22] caTools_1.17.1                memoise_1.1.0                 biomaRt_2.32.1               
[25] httpuv_1.3.5                  curl_2.7                      Rcpp_0.12.11                 
[28] KernSmooth_2.23-15            xtable_1.8-2                  gdata_2.18.0                 
[31] DelayedArray_0.2.7            XVector_0.16.0                mime_0.5                     
[34] bit_1.1-12                    Rsamtools_1.28.0              AnnotationHub_2.8.2          
[37] digest_0.6.12                 stringi_1.1.5                 shiny_1.0.3                  
[40] grid_3.4.1                    tools_3.4.1                   bitops_1.0-6                 
[43] magrittr_1.5                  RCurl_1.95-4.8                lazyeval_0.2.0               
[46] tibble_1.3.3                  RSQLite_2.0                   pkgconfig_2.0.1              
[49] Matrix_1.2-10                 httr_1.2.1                    R6_2.2.2                     
[52] GenomicAlignments_1.12.1      compiler_3.4.1 
ADD COMMENTlink modified 16 days ago by Gordon Smyth31k • written 16 days ago by sarah.williams120
1
gravatar for Gordon Smyth
16 days ago by
Gordon Smyth31k
Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia
Gordon Smyth31k wrote:

Hi Sarah,

Yes, you're quite right. Thanks for reporting it. I've committed a fix now to limma 3.32.6.

Reporting a bug to the support site like this is fine, but I encourage people to email me bug reports directly -- see Section 2.3 of the limma User's Guide.

ADD COMMENTlink modified 14 days ago • written 16 days ago by Gordon Smyth31k
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: 104 users visited in the last hour