Re: [MORPHMET] determining if groups of specimens are statistically distinct.

[Dean Adams]

Ian,



The command summary(manova(..)) will perform a parametric manova. However,
for the datafile you sent, that operation will not be possible because the
number of variables relative to sample size (note, you posted a similar
issue about a month ago to Morphmet).


The alternative is to use permutation-based evaluation of group
differences. Permutation-based manova can be performed in numerous
R-packages, including geomorph and RRPP.  These two packages also have the
ability to perform pairwise comparisons using the permutation regime
specified.



Example R-code for implementing both is found below (NOTE: I strongly
suggest you read the help files and their examples to better understand the
options for these functions: there are many).


Best,


Dean



##

mydat <- read.csv('PCWIDE1.txt', header = TRUE,sep = '\t')

spec <- as.factor(mydat[,1])

shape <- as.matrix(mydat[,-1])



summary(manova(shape~spec))  #won’t work.  df issue related to N vs. P



library(geomorph)

gdf<-geomorph.data.frame(spec=spec,shape=shape)

procD.lm(shape~spec,data=gdf)

advanced.procD.lm(shape~spec,f2 = ~1, group = ~spec, data=gdf)



library(RRPP)

rdf<-rrpp.data.frame(spec=spec,shape=shape)

res <- lm.rrpp(shape~spec,data=rdf)

res.null<- lm.rrpp(shape~1,data=rdf)

anova(res)

PW <- pairwise(fit=res, groups = spec)

summary(PW)

On Thu, Oct 4, 2018 at 2:06 PM Ian  wrote:

> I am trying to determine if three groups are statistically distinct from
> one another. As is my script produces P values for each x and y coordinate
> but i am unsure how to get a comparison between the groups as a whole. My
> script and example data are attached. Can anyone tell me how i would go
> about determine if the groups Rafalt, Rafnas, and Rafpon are statistically
> distinct? I have tried to make subsets to do a pairwise comparison but that
> has not been successful. Any input is much appreciated.
>
>
> data1<-PCWIDE1
>
> names(data1)
>
> Raf.manova1<-
> manova(cbind(x1,y1,x2,y2,x3,y3,x4,y4,x5,y5,x6,y6,x7,y7,x8,y8,x9,y9,x10,y10,x11,y11,x12,y12,x13,y13,x14,y14,x15,y15,x16,y16,x17,y17,x18,y18,x20,y20,x21,y21,x22,y22,x23,y23,x24,y24,x25,y25,x26,y26,x27,y27,x28,y28,x29,y29,x30,y30,x31,y31)~
> as.factor(Species),data=data1)
>
> SAV<-summary.aov(Raf.manova1)
>
> print(SAV)
>
> --
> MORPHMET may be accessed via its webpage at http://www.morphometrics.org
> ---
> You received this message because you are subscribed to the Google Groups
> "MORPHMET" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to morphmet+unsubscr...@morphometrics.org.
>


-- 

Dr. Dean C. Adams

Director of Graduate Education, EEB Program

Professor

Department of Ecology, Evolution, and Organismal Biology

Iowa State University

-- 
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
--- 
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.

[MORPHMET] PLS analysis skull ontogeny MorphoJ: weird pattern observed

[Agnese Lanzetti]

Hi all,

I am getting weird results when doing a PLS analysis in MorphoJ.

I have a dataset of 12 minke whale skulls of different ontogenetic ages and I 
wanted to test if the anterior part (rostrum) and posterior part (braincase) 
are modular and how their shape changes relative to one another during growth.
Both parts are blocks of the same configuration and they are on the same 
Procrustes fit, as recommended for these kinds of studies.

I am getting strange results on the block 1 PLS1 vs block 2 PLS1 graph, with 
all the specimens falling on the same "line" (looks like a linear regression 
with a 45' slope), but instead of having the specimens ordered from the 
youngest to the oldest, with youngest close to 0 and oldest at the other end, 
the pattern is strange.
The adults fall around the middle of the distribution, with the youngest 
specimens close to 0 and intermediate specimens at the top of the distribution. 
So it kind of seems that the shapes are "reverting" back in adults compared to 
younger specimens.

Unfortunately I can't attach an image so I tried to describe the problem in the 
best way possible.
Have you ever seen a pattern as the one I described? If not, in your experience 
is it biologically possible or it implies that there is something wrong with my 
data?
I think my limited sample size might be part of the problem, but I'm not sure 
if that's the only reason why I'm seeing this pattern.

Let me know if you need additional info.
Thank you very much!
Best,
Agnese Lanzetti

agnese.lanze...@hotmail.it
Ph.D. candidate
Joint Doctoral Program in Evolutionary Biology
San Diego State University
University of California, Riverside

-- 
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
--- 
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.