[BioC] error in DEXSeq fitDispersionFunction()

[Georg Otto]

Wolfgang Huber <whuber at embl.de> writes:

Dear Wolfgang,

thanks for your help. I tried with version 1.1.1 and still got the same
message:

Error in if (sum(log(coefs/oldcoefs)^2) < 0.005) break : 
  missing value where TRUE/FALSE needed
In addition: Warning messages:
1: In glmgam.fit(mm, disps[good], start = coefs) :
  Too much damping - convergence tolerance not achievable
2: In log(coefs/oldcoefs) : NaNs produced

Apparently in the function fitDispersionFunction(), the expression

sum(log(coefs/oldcoefs)^2)

produces NAs with my data, so 

if (sum(log(coefs/oldcoefs)^2) < 0.005)

gives an error... any idea, how to fix this?

Cheers,

Georg


> Dear Georg
>
> Thank you. Can you please try with a more recent version of DEXSeq,
> e.g. 1.1.1 from the last release? The package has changed a lot since
> 1.0.1, including in places that appear relevant to your problem.
>
> 	Best wishes
> 	Wolfgang
>
> 	
>
> On 11/29/11 3:30 PM, Georg Otto wrote:
>> Dear Bioconductors,
>>
>>
>> I am trying to use DEXSeq to test for differential exon usage in my
>> RNA-seq data. I was able to generate the exon count tables using the
>> python scripts that come with the package and the function
>> read.HTSeqCounts(). However, I get an error at this step:
>>
>> exon.counts<- fitDispersionFunction(exon.counts)
>>
>>   Error in if (sum(log(coefs/oldcoefs)^2)<  0.005) break :
>>     missing value where TRUE/FALSE needed
>>   In addition: Warning messages:
>>   1: In glmgam.fit(mm, disps[good], start = coefs) :
>>     Too much damping - convergence tolerance not achievable
>>   2: In log(coefs/oldcoefs) : NaNs produced
>>
>> Any idea what might be wrong here?
>>
>> Here is the output of sessionInfo()
>>
>>
>> R version 2.13.0 (2011-04-13)
>> Platform: x86_64-pc-linux-gnu (64-bit)
>>
>> locale:
>>   [1] LC_CTYPE=en_GB.UTF-8       LC_NUMERIC=C
>>   [3] LC_TIME=en_GB.UTF-8        LC_COLLATE=en_GB.UTF-8
>>   [5] LC_MONETARY=C              LC_MESSAGES=en_GB.UTF-8
>>   [7] LC_PAPER=en_GB.UTF-8       LC_NAME=C
>>   [9] LC_ADDRESS=C               LC_TELEPHONE=C
>> [11] LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C
>>
>> attached base packages:
>> [1] stats     graphics  grDevices utils     datasets  methods   base
>>
>> other attached packages:
>> [1] DEXSeq_1.0.1          DESeq_1.6.1           locfit_1.5-6
>> [4] lattice_0.19-26       akima_0.5-4           Biobase_2.14.0
>> [7] GenomicFeatures_1.4.4 GenomicRanges_1.4.8   IRanges_1.10.6
>>
>> loaded via a namespace (and not attached):
>>   [1] annotate_1.30.0      AnnotationDbi_1.14.1 biomaRt_2.8.1
>>   [4] Biostrings_2.20.1    BSgenome_1.20.0      DBI_0.2-5
>>   [7] genefilter_1.34.0    geneplotter_1.30.0   grid_2.13.0
>> [10] hwriter_1.3          plyr_1.6             RColorBrewer_1.0-2
>> [13] RCurl_1.7-0          RSQLite_0.10.0       rtracklayer_1.12.4
>> [16] splines_2.13.0       statmod_1.4.14       stringr_0.5
>> [19] survival_2.36-9      tools_2.13.0         XML_3.4-3
>> [22] xtable_1.6-0
>>
>>
>>
>> Cheers,
>>
>> Georg
>>
>> _______________________________________________
>> Bioconductor mailing list
>> Bioconductor at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor