R help - Sep 2015 - How to find out if two cells in a dataframe belong to the same pre-specified factor-level

Dear list,
I really couldnt find a better way to describe my question, so please  
bear with me.

To illustrate my problem, i have a matrix with ecological distances  
(m1) and one with genetic distances (m2) for a number of biological  
species. I have merged both matrices and want to plot both distances  
versus each other, as illustrated in this example:

library(reshape)
library(ggplot2)
library(dplyr)

dist1 <- matrix(runif(16),4,4)
dist2 <- matrix(runif(16),4,4)
rownames(dist1) <- colnames(dist1) <- paste0("A",1:4)
rownames(dist2) <- colnames(dist2) <- paste0("A",1:4)

m1 <- melt(dist1)
m2 <- melt(dist2)

final <- full_join(m1,m2, by=c("Var1","Var2"))
ggplot(final, aes(value.x,value.y)) + geom_point()

Here is the twist:
The biological species belong to certain groups, which are given in  
the dataframe `species`, for example:

species <- data.frame(spcs=as.character(paste0("A",1:4)),
                      
grps=as.factor(c(rep("cat",2),(rep("dog",2)))))

I want to check if a x,y pair in final (as in `final$Var1`,  
`final$Var2`) belongs to the same group of species (here "cat" or  
"dog"), and then want to color all groups specifically in the  
x,y-scatterplot.
Thus, i need an R translation for:

final$group <- If (final$Var1 and final$Var2) belong to the same group  
as specified
       in species, then assign the species group here, else do nothing  
or assign NA

so i can proceed with

ggplot(final, aes(value.x,value.y, col=group)) + geom_point()

So, in the example, the pairs A1-A1, A1-A2, A2-A1, A2-A2 should be  
identified as "both cats", hence should get the factor
"cat".

Thank you very much!


Tim

Here's one approach that works.  I made some changes to the code you
provided.  Full working example code given below.

library(reshape)
library(ggplot2)
library(dplyr)

dist1 <- matrix(runif(16), 4, 4)
dist2 <- matrix(runif(16), 4, 4)
rownames(dist1) <- colnames(dist1) <- paste0("A", 1:4)
rownames(dist2) <- colnames(dist2) <- paste0("A", 1:4)
m1 <- melt(dist1)
m2 <- melt(dist2)
# I changed the by= argument here
final <- full_join(m1, m2, by=c("X1", "X2"))

# I made some changes to keep spcs character and grps factor
species <- data.frame(spcs=paste0("A", 1:4),
  grps=as.factor(c(rep("cat", 2), (rep("dog", 2)))),
stringsAsFactors=FALSE)

# define new variables for final indicating group membership
final$g1 <- species$grps[match(final$X1, species$spcs)]
final$g2 <- species$grps[match(final$X2, species$spcs)]
final$group <- as.factor(with(final, ifelse(g1==g2, as.character(g1),
"dif")))

# plot just the rows with matching groups
ggplot(final[final$group!="dif", ], aes(value.x, value.y, col=group))
+
  geom_point()
# plot all the rows
ggplot(final, aes(value.x, value.y, col=group)) + geom_point()

Jean


On Sun, Sep 27, 2015 at 4:22 PM, <trichter at uni-bremen.de> wrote:
> Dear list,
> I really couldnt find a better way to describe my question, so please bear
> with me.
>
> To illustrate my problem, i have a matrix with ecological distances (m1)
> and one with genetic distances (m2) for a number of biological species. I
> have merged both matrices and want to plot both distances versus each
> other, as illustrated in this example:
>
> library(reshape)
> library(ggplot2)
> library(dplyr)
>
> dist1 <- matrix(runif(16),4,4)
> dist2 <- matrix(runif(16),4,4)
> rownames(dist1) <- colnames(dist1) <- paste0("A",1:4)
> rownames(dist2) <- colnames(dist2) <- paste0("A",1:4)
>
> m1 <- melt(dist1)
> m2 <- melt(dist2)
>
> final <- full_join(m1,m2, by=c("Var1","Var2"))
> ggplot(final, aes(value.x,value.y)) + geom_point()
>
> Here is the twist:
> The biological species belong to certain groups, which are given in the
> dataframe `species`, for example:
>
> species <- data.frame(spcs=as.character(paste0("A",1:4)),
>                      
grps=as.factor(c(rep("cat",2),(rep("dog",2)))))
>
> I want to check if a x,y pair in final (as in `final$Var1`, `final$Var2`)
> belongs to the same group of species (here "cat" or
"dog"), and then want
> to color all groups specifically in the x,y-scatterplot.
> Thus, i need an R translation for:
>
> final$group <- If (final$Var1 and final$Var2) belong to the same group
as
> specified
>       in species, then assign the species group here, else do nothing or
> assign NA
>
> so i can proceed with
>
> ggplot(final, aes(value.x,value.y, col=group)) + geom_point()
>
> So, in the example, the pairs A1-A1, A1-A2, A2-A1, A2-A2 should be
> identified as "both cats", hence should get the factor
"cat".
>
> Thank you very much!
>
>
> Tim
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> 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.
>
	[[alternative HTML version deleted]]

Thank you, that turned out to work very well.

If you want to, you can answer it here:

http://stackoverflow.com/questions/32809249/how-to-find-out-if-two-cells-in-a-dataframe-belong-to-the-same-pre-specified-fac/

The question wasnt properly answered, which is why i switched to R-list.


On 28.09.2015 20:15, Adams, Jean wrote:
> Here's one approach that works.  I made some changes to the code you 
> provided.  Full working example code given below.
>
> library(reshape)
> library(ggplot2)
> library(dplyr)
>
> dist1 <- matrix(runif(16), 4, 4)
> dist2 <- matrix(runif(16), 4, 4)
> rownames(dist1) <- colnames(dist1) <- paste0("A", 1:4)
> rownames(dist2) <- colnames(dist2) <- paste0("A", 1:4)
> m1 <- melt(dist1)
> m2 <- melt(dist2)
> # I changed the by= argument here
> final <- full_join(m1, m2, by=c("X1", "X2"))
>
> # I made some changes to keep spcs character and grps factor
> species <- data.frame(spcs=paste0("A", 1:4),
>   grps=as.factor(c(rep("cat", 2), (rep("dog", 2)))), 
> stringsAsFactors=FALSE)
>
> # define new variables for final indicating group membership
> final$g1 <- species$grps[match(final$X1, species$spcs)]
> final$g2 <- species$grps[match(final$X2, species$spcs)]
> final$group <- as.factor(with(final, ifelse(g1==g2, as.character(g1), 
> "dif")))
>
> # plot just the rows with matching groups
> ggplot(final[final$group!="dif", ], aes(value.x, value.y,
col=group)) +
>   geom_point()
> # plot all the rows
> ggplot(final, aes(value.x, value.y, col=group)) + geom_point()
>
> Jean
>
>
> On Sun, Sep 27, 2015 at 4:22 PM, <trichter at uni-bremen.de 
> <mailto:trichter at uni-bremen.de>> wrote:
>
>     Dear list,
>     I really couldnt find a better way to describe my question, so
>     please bear with me.
>
>     To illustrate my problem, i have a matrix with ecological
>     distances (m1) and one with genetic distances (m2) for a number of
>     biological species. I have merged both matrices and want to plot
>     both distances versus each other, as illustrated in this example:
>
>     library(reshape)
>     library(ggplot2)
>     library(dplyr)
>
>     dist1 <- matrix(runif(16),4,4)
>     dist2 <- matrix(runif(16),4,4)
>     rownames(dist1) <- colnames(dist1) <- paste0("A",1:4)
>     rownames(dist2) <- colnames(dist2) <- paste0("A",1:4)
>
>     m1 <- melt(dist1)
>     m2 <- melt(dist2)
>
>     final <- full_join(m1,m2, by=c("Var1","Var2"))
>     ggplot(final, aes(value.x,value.y)) + geom_point()
>
>     Here is the twist:
>     The biological species belong to certain groups, which are given
>     in the dataframe `species`, for example:
>
>     species <- data.frame(spcs=as.character(paste0("A",1:4)),
>     grps=as.factor(c(rep("cat",2),(rep("dog",2)))))
>
>     I want to check if a x,y pair in final (as in `final$Var1`,
>     `final$Var2`) belongs to the same group of species (here
"cat" or
>     "dog"), and then want to color all groups specifically in the
>     x,y-scatterplot.
>     Thus, i need an R translation for:
>
>     final$group <- If (final$Var1 and final$Var2) belong to the same
>     group as specified
>           in species, then assign the species group here, else do
>     nothing or assign NA
>
>     so i can proceed with
>
>     ggplot(final, aes(value.x,value.y, col=group)) + geom_point()
>
>     So, in the example, the pairs A1-A1, A1-A2, A2-A1, A2-A2 should be
>     identified as "both cats", hence should get the factor
"cat".
>
>     Thank you very much!
>
>
>     Tim
>
>     ______________________________________________
>     R-help at r-project.org <mailto:R-help at r-project.org> mailing
list --
>     To UNSUBSCRIBE and more, see
>     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.
>
>

-- 
Tim Richter-Heitmann (M.Sc.)
PhD Candidate



International Max-Planck Research School for Marine Microbiology
University of Bremen
Microbial Ecophysiology Group (AG Friedrich)
FB02 - Biologie/Chemie
Leobener Stra?e (NW2 A2130)
D-28359 Bremen
Tel.: 0049(0)421 218-63062
Fax: 0049(0)421 218-63069