Hey,
I have a difficult question(for me at least).
I did the sam calculation and got the data i wanted. First i have a
matrix with
double values (last.out)
>last.out<- sam(last.data, AE0625419.cl, B=20,
gene.names=matrix_all_names[,2])
After i cut out the lines with 'NA'-values for the p-values
> scores <- log(pmax(last.out at p.value))
> valids <- !is.na(scores)
> scores <- scores[valids]
> data <- last.data[valids,]
I have a list of data (18 Columns x 19,302 rows)
I then added the p-values as a 19th column to the list
> data[,19] <- scores
Afterwards I've searched for the significant genes (delta =10)
> sig <- list.siggenes(last.out, 10)
subsequent I deduced the data matrix to the few genes which were
significant.
> data_end <- last.data[sig,]
Now I come to my problem(s):
1. I would like to have my p-Values also in the last data matrix.
Somehow I lose
these values while minimizing the matrix.
the data file is than again with only 18 columns.
>dim(data_end)
[1] 6 18
But if i add the scores afterwards then the list is too long, because
i have
only 6 rows (and not 19,302).
Can you help me?
2. My second problem is a bit bigger.
I would like to convert my first data file (data) into a logical
matrix, with
TRUE for the rows which are in the object sig( the significantly
expressed
genes) The other rows should be FALSE
I nthis case it should look like a matrix with 19,302 rows, whereas 6
rows have
TRUE in all 18 columns and all the rest have FALSE ( I would be happy
to do it
also with 0=FALSE and 1=TRUE, but I then have the same problem.)
It would be really nice if someone can help me with this two cases.
TVM,
Assa
--
Assa Yeroslaviz
L?tzenerstr. 15
51373 Leverkusen
On 11/4/05 4:32 AM, "Assa Yeroslaviz" <assa.yeroslaviz at="" uni-="" duesseldorf.de="">
wrote:
> Hey,
>
> I have a difficult question(for me at least).
>
> I did the sam calculation and got the data i wanted. First i have a
matrix
> with
> double values (last.out)
>> last.out<- sam(last.data, AE0625419.cl, B=20,
>> gene.names=matrix_all_names[,2])
>
> After i cut out the lines with 'NA'-values for the p-values
>> scores <- log(pmax(last.out at p.value))
>> valids <- !is.na(scores)
>> scores <- scores[valids]
>> data <- last.data[valids,]
>
> I have a list of data (18 Columns x 19,302 rows)
> I then added the p-values as a 19th column to the list
>> data[,19] <- scores
>
> Afterwards I've searched for the significant genes (delta =10)
>> sig <- list.siggenes(last.out, 10)
>
> subsequent I deduced the data matrix to the few genes which were
significant.
>> data_end <- last.data[sig,]
>
> Now I come to my problem(s):
>
> 1. I would like to have my p-Values also in the last data matrix.
Somehow I
> lose
> these values while minimizing the matrix.
> the data file is than again with only 18 columns.
>> dim(data_end)
> [1] 6 18
>
> But if i add the scores afterwards then the list is too long,
because i have
> only 6 rows (and not 19,302).
>
> Can you help me?
If the order of what you give above is correct, then "scores" was
never part
of last.data and there is no reason to expect that indexing last.data
is
going to somehow include "scores"--they were haver added. If I
understand
what you are doing and want to do, just do:
scores <- log(pmax(last.out at p.value))
data <- cbind(last.data,scores)
data.end <- data[sig,]
> 2. My second problem is a bit bigger.
>
> I would like to convert my first data file (data) into a logical
matrix, with
> TRUE for the rows which are in the object sig( the significantly
expressed
> genes) The other rows should be FALSE
> I nthis case it should look like a matrix with 19,302 rows, whereas
6 rows
> have
> TRUE in all 18 columns and all the rest have FALSE ( I would be
happy to do it
> also with 0=FALSE and 1=TRUE, but I then have the same problem.)
>
> It would be really nice if someone can help me with this two cases.
mynewmatrix <- matrix(FALSE,nr=19302,nc=18)
mynewmatrix[sig,] <- TRUE
Just out of curiosity, why do you want to do this? It contains no
more
information than sig by itself?
Sean