[BioC] question on "uniCox"

Vincent Carey stvjc at channing.harvard.edu
Sun Nov 21 10:19:41 CET 2010


uniCox is not a bioconductor package.  the question would probably do
better on r-help, or you could contact the uniCox maintainer.

On Sun, Nov 21, 2010 at 9:51 AM, Shi, Tao <shidaxia at yahoo.com> wrote:
> Hi list,
>
> I’m testing out uniCox R package (version 1.0, on R2.12.0, WinXP).
>
> When I ran uniCox on my data, there are always some NA’s in the beta matrix,
> which in turn causes problems in uniCoxCV call.  I don’t see anything  wrong
> with the corresponding data (e.g. no NAs) and if I fit a  univariate Cox model,
> the features that give NA beta estimates are  actually pretty significant.
> Could you please let me know what  happened and how to avoid this?
>
> I’ve attached the outputs of the function calls below.
>
> Thank you very much!
>
>
> ...Tao
>
>
>> a <- uniCox(x=t(dat.ave.train.base), y=sampleinfo.ave.train.base$tm2dthr,
>>status=sampleinfo.ave.train.base$censrdth)
> lambda value  1 out of  20
> lambda value  2 out of  20
> lambda value  3 out of  20
> lambda value  4 out of  20
> lambda value  5 out of  20
> lambda value  6 out of  20
> lambda value  7 out of  20
> lambda value  8 out of  20
> lambda value  9 out of  20
> lambda value  10 out of  20
> lambda value  11 out of  20
> lambda value  12 out of  20
> lambda value  13 out of  20
> lambda value  14 out of  20
> lambda value  15 out of  20
> lambda value  16 out of  20
> lambda value  17 out of  20
> lambda value  18 out of  20
> lambda value  19 out of  20
> lambda value  20 out of  20
> 5  betas missing
>
>>  aa <- uniCoxCV(a, x=t(dat.ave.train.base),
>>y=sampleinfo.ave.train.base$tm2dthr,  status=sampleinfo.ave.train.base$censrdth)
> FOLD= 1
> lambda value  1 out of  20
> lambda value  2 out of  20
> lambda value  3 out of  20
> lambda value  4 out of  20
> lambda value  5 out of  20
> lambda value  6 out of  20
> lambda value  7 out of  20
> lambda value  8 out of  20
> lambda value  9 out of  20
> lambda value  10 out of  20
> lambda value  11 out of  20
> lambda value  12 out of  20
> lambda value  13 out of  20
> lambda value  14 out of  20
> lambda value  15 out of  20
> lambda value  16 out of  20
> lambda value  17 out of  20
> lambda value  18 out of  20
> lambda value  19 out of  20
> lambda value  20 out of  20
> 3  betas missing
> 1
> Error in coxph(Surv(y[ii], status[ii]) ~ eta.new) :
>  No (non-missing) observations
>
>> a[[2]][(rowSums(is.na(a[[2]])))>0,]
>         [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10] [,11] [,12] [,13]
> [,14] [,15] [,16] [,17] [,18] [,19] [,20]
> [1,]  92.6641  NaN  NaN  NaN    0    0    0    0    0     0     0     0     0
> 0     0     0     0     0     0     0
> [2,]      NaN    0    0    0    0    0    0    0    0     0     0     0     0
> 0     0     0     0     0     0     0
> [3,] 567.3650  NaN    0    0    0    0    0    0    0     0     0     0     0
> 0     0     0     0     0     0     0
>
>
>
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor



More information about the Bioconductor mailing list