[BioC] question on "uniCox"

[Shi, Tao]

Thanks, Vincent!  I'm aware of that, but just thought this is probably more 
relevant group.  I'll try r-help too.

...Tao



----- Original Message ----
> From: Vincent Carey <stvjc at channing.harvard.edu>
> To: "Shi, Tao" <shidaxia at yahoo.com>
> Cc: BioC <bioconductor at stat.math.ethz.ch>
> Sent: Sun, November 21, 2010 1:19:41 AM
> Subject: Re: [BioC] question on "uniCox"
> 
> 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
>