faster execution of for loop in Fishers test

On 12 Feb 2019, at 02:45 , Bert Gunter <bgunter.4567 at gmail.com> wrote:

1. I believe Fisher's exact test is computationally intensive and takes a
lot of time for large structures, so I would say what you see is what you
should expect! (As I'm not an expert on this, confirmation or contradiction
by those who are would be appreciated).

2. Your second question on how to select results based on values in another
vector/column is very basic R. So it appears that you need to spend some
time with an R tutorial or two to learn the basics (unless I have
misinterpreted).

3. Please do not repost further. No one is obligated to respond to your
posts. Following the posting guide, which you appear to have done,
increases the likelihood, but is of course no guarantee.

Cheers,
Bert


Bert Gunter

"The trouble with having an open mind is that people keep coming along and
sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )


On Mon, Feb 11, 2019 at 5:28 PM Adrian Johnson <oriolebaltimore at gmail.com>
wrote:

Dear group,

I have two large matrices.

Matrix one: is 24776 x 76 (example toy1 dput object given below)

Matrix two: is 12913 x 76 (example toy2 dput object given below)

Column names of both matrices are identical.

My aim is:

a. Take each row of toy2 and transform vector into UP (>0)  and DN (
<0 ) categories. (kc)
b  Test association between kc and every row of toy1.

My code, given below, although this works but is very slow.

I gave dput objects for toy1, toy2 and result matrix.

Could you suggest/help me how I can make this faster.  Also, how can I
select values in result column that are less than 0.001 (p < 0.001).

Appreciate your help. Thank you.
-Adrian

Code:

===============================================================================



result <- matrix(NA,nrow=nrow(toy1),ncol=nrow(toy2))

rownames(result) <- rownames(toy1)
colnames(result) <- rownames(toy2)

for(i in 1:nrow(toy2)){
for(j in 1:nrow(toy1)){
kx = toy2[i,]
kc <- rep('NC',length(kx))
kc[ kx >0] <- 'UP'
kc[ kx <=0 ] <- 'DN'
xpv <- fisher.test(table(kc,toy1[j,]),simulate.p.value = TRUE)$p.value
result[j,i] <- xpv
}
}


===============================================================================



===============================================================================

dput(toy1)

structure(c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1), .Dim = c(10L, 7L), .Dimnames = list(c("ACAP3",
"ACTRT2", "AGRN", "ANKRD65", "ATAD3A", "ATAD3B", "ATAD3C", "AURKAIP1",
"B3GALT6", "C1orf159"), c("a", "b", "c", "d", "e", "f", "g")))

dput(toy2)

structure(c(-0.242891119688613, -0.0514058216682132, 0.138447212993773,
-0.312576648033122, 0.271489918720452, -0.281196468299486,
-0.0407160143344565,
-0.328353812845287, 0.151667836674511, 0.408596843743938,
-0.049351944902924,
0.238586287349249, 0.200571558784821, -0.0737604184858411,
0.245971526254877,
0.24740263959845, -0.161528943131908, 0.197521973013793,
0.0402668125708444,
0.376323735212088, 0.0731550871764204, 0.385270176969893, 0.28953042756208,
0.062587289401188, -0.281187168932979, -0.0202298984561554,
-0.0848696970309447,
0.0349676726358973, -0.520484215644868, -0.481991414222996,
-0.00698099201388211,
0.135503878341873, 0.156983081312087, 0.320223832092661, 0.34582193394074,
0.0844455960468667, -0.157825604090972, 0.204758250510969,
0.261796072978612,
-0.19510450641405, 0.43196474472874, -0.211155577453175,
-0.0921641871215187,
0.420950361292263, 0.390261862151936, -0.422273930504427,
0.344653684951627,
0.0378273248838503, 0.197782027324611, 0.0963124876309569,
0.332093167080656,
0.128036554821915, -0.41338065859335, -0.409470440033177,
0.371490567256253,
-0.0912549189140141, -0.247451812684234, 0.127741739114639,
0.0856254238844557,
0.515282940316031, -0.25675759521248, 0.333943163209869, 0.604141413840881,
0.0824942299510931, -0.179605710473021, -0.275604207054643,
-0.113251154591898,
0.172897837449258, -0.329808795076691, -0.239255324324506), .Dim = c(10L,
7L), .Dimnames = list(c("chr5q23", "chr16q24", "chr8q24", "chr13q11",
"chr7p21", "chr10q23", "chr13q13", "chr10q21", "chr1p13", "chrxp21"
), c("a", "b", "c", "d", "e", "f", "g")))

dput(result)

structure(c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0.532733633183408,
0.511244377811094, 0.528235882058971, 0.526736631684158, 0.51424287856072,
0.530734632683658, 0.513243378310845, 0.533233383308346, 0.542228885557221,
0.517241379310345, 0.532733633183408, 0.521739130434783, 0.529235382308846,
0.530234882558721, 0.548725637181409, 0.525737131434283, 0.527236381809095,
0.532733633183408, 0.530234882558721, 0.520739630184908, 0.15592203898051,
0.142928535732134, 0.140929535232384, 0.150924537731134, 0.160419790104948,
0.139430284857571, 0.152923538230885, 0.146426786606697, 0.149425287356322,
0.145427286356822, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0.282358820589705,
0.293853073463268, 0.262868565717141, 0.290854572713643, 0.276861569215392,
0.288855572213893, 0.282358820589705, 0.292853573213393, 0.286356821589205,
0.271364317841079, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1), .Dim = c(10L, 10L), .Dimnames = list(c("ACAP3",
"ACTRT2", "AGRN", "ANKRD65", "ATAD3A", "ATAD3B", "ATAD3C", "AURKAIP1",
"B3GALT6", "C1orf159"), c("chr5q23", "chr16q24", "chr8q24", "chr13q11",
"chr7p21", "chr10q23", "chr13q13", "chr10q21", "chr1p13", "chrxp21"
)))

	[[alternative HTML version deleted]]