R help - Nov 2009 - when vectorising does not work: silent function fail?

Dear All,

I'm using apply to do some genetic association analysis along a chromosome,
with
many thousands markers. For each marker the analysis is the same, so I was 
planning to use apply(chrom, 2, somefunction)

In the specific case I do:

my.results = apply(chr, 2, function(x){anova(lrm( cpstc.f ~ x + time.cpstc + age
+ sex + mri))[1,3]})

This is all good and well in theory, but in practice the lrm() model will fail 
for some markers and wreck the whole process. Failure for some markers is no 
problem for me, but *predicting* which markers will fail can be hugely
problematic.

I then though of creating some fucntion to catch the error messages that would 
otherwise scr*w things over:

my.lrm = function(x){
pol = NULL
pol = lrm( cpstc.f ~ x + time.cpstc + age + sex + mri)
if(length(pol) > 0)
rez = anova(pol)[1,3]
if(length(pol)  == 0)
rez = 1
rez}

my.results = apply(chr, 2, my.lrm)

Still no joy, even adding try() in the evaluation and 
options(show.error.messages = F)

I am at loss on how to get the darn function to bail out *silently* if needs be 
so I can just smack a replacement value in --which would also have the benefit 
of keeping the order of the markers.

Any idea will be gratefully acknowledged.

Best,

Federico


-- 
Federico C. F. Calboli
Department of Epidemiology and Public Health
Imperial College, St Mary's Campus
Norfolk Place, London W2 1PG

Tel  +44 (0)20 7594 1602     Fax (+44) 020 7594 3193

f.calboli [.a.t] imperial.ac.uk
f.calboli [.a.t] gmail.com

Have you tried something like this:

my.results = apply(chr, 2, function(x){
    result <- try(anova(lrm( cpstc.f ~ x + time.cpstc + age + sex +
mri))[1,3])
    if (inherits(result, "try-error")) return(NULL)
    result
})

This should catch the error and have NULL in that list element.

On Tue, Nov 10, 2009 at 10:04 AM, Federico Calboli
<f.calboli at imperial.ac.uk> wrote:
> Dear All,
>
> I'm using apply to do some genetic association analysis along a
chromosome,
> with many thousands markers. For each marker the analysis is the same, so I
> was planning to use apply(chrom, 2, somefunction)
>
> In the specific case I do:
>
> my.results = apply(chr, 2, function(x){anova(lrm( cpstc.f ~ x + time.cpstc
+
> age + sex + mri))[1,3]})
>
> This is all good and well in theory, but in practice the lrm() model will
> fail for some markers and wreck the whole process. Failure for some markers
> is no problem for me, but *predicting* which markers will fail can be
hugely
> problematic.
>
> I then though of creating some fucntion to catch the error messages that
> would otherwise scr*w things over:
>
> my.lrm = function(x){
> pol = NULL
> pol = lrm( cpstc.f ~ x + time.cpstc + age + sex + mri)
> if(length(pol) > 0)
> rez = anova(pol)[1,3]
> if(length(pol) ?== 0)
> rez = 1
> rez}
>
> my.results = apply(chr, 2, my.lrm)
>
> Still no joy, even adding try() in the evaluation and
> options(show.error.messages = F)
>
> I am at loss on how to get the darn function to bail out *silently* if
needs
> be so I can just smack a replacement value in --which would also have the
> benefit of keeping the order of the markers.
>
> Any idea will be gratefully acknowledged.
>
> Best,
>
> Federico
>
>
> --
> Federico C. F. Calboli
> Department of Epidemiology and Public Health
> Imperial College, St Mary's Campus
> Norfolk Place, London W2 1PG
>
> Tel ?+44 (0)20 7594 1602 ? ? Fax (+44) 020 7594 3193
>
> f.calboli [.a.t] imperial.ac.uk
> f.calboli [.a.t] gmail.com
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>


-- 
Jim Holtman
Cincinnati, OH
+1 513 646 9390

What is the problem that you are trying to solve?