Post date: May 22, 2017 4:22:16 PM
I'm using MorphoJ 1.06d.
23v17. size_and_coord-raw.csv just has all the raw measurements, size and position. size-raw is just the raw size meas, coord-raw is just the raw position meas. To be read by MorphoJ, you have to get rid of the pop column, and replace NA with -9999.
size.morphoj:
Uploaded raw size data > Preliminaries
>Procrustes fit (align by principal axes).
>Extract New Classifier From ID String
But, hello, MorphoJ is for the position data, not the size data. So...
coord.morphoj:
When getting this file to upload, I found that CSP12-077rhf has 19 area measurements, not 18, so I'll have Megan remeasure this one and I'll drop it out for now. See others in To_remeasure_23may17.xlsx.
coord-raw-nowing.txt is the file with xy coord for the entire wing removed. This is the file I uploaded to MorphoJ. There were a couple "N"'s I turned into -9999's (NA typos). There was a randomly inserted "-9999" in a cell, no idea why.
>Extract New Classifier From ID String
>Procrustes fit (align by principal axes)
>Find Outliers...
The following dropped out due to missing data (NA -- wing was ripped):
CDY12-003
DBS10-001
DBS10-011
DBS10-023
DBS10-024
DOP02-002
FCR12-017
GNP08-019
KHL09-021
MBM07-200
MTR12-058
SIN10-007
SIN10-026
SIN10-045
SIN10-046
SIN10-074
SIN11-010
SIN11-021
SOP11-008
SOP11-010
SOP11-018
YWP12-025
YWP12-039
These dropped out because there was no spot on the wing, meaning there's no position; a 0 in the dataset (they will not need to drop out of the area dataset??):
ABM12-003
ABM12-038
ABM12-044
ABM12-020
ABM12-028
ABM12-024
ABM12-033
ABM12-011
ABM12-008
ABM12-015
ABM12-019
ABM12-023
ABM12-039
ABM12-045
ABM12-006
BIG12-003
BIG12-006
BIG12-007
BIG12-034
BIG12-002
BIG12-022
BIG12-031
BKM12-008
BKM12-055
BKM12-005
GNP12-036
GNP08-028
LKS12-003
MBM07-201
MBM07-210
MBM07-206
MBM07-204
MBM07-207
SHC11-020
SHC11-039
SHC11-047
SHC11-040
SHC11-016
SIN11-020
SIN10-034
SLA12-050
YBG11-008
YBG12-057
YBG11-012
YBG11-015
YBG11-031
YBG12-003
YBG12-031
YBG12-038
YBG12-044
YBG12-067
YBG12-079
YBG12-080
YBG12-088
YBG12-091
YBG12-092
YBG12-094
YBG12-096
and others!! turn 0's into -9999's in this dataset.
I played more with the dataset even though some things need fixed. I added a wireframe (connected the margins (18 through 23 connect) and veins (18 to 24, 19 to 24, 21 to 25 and 26, 22 to 25, 23 to 26) together), I made a covariance matrix, then did a PCA. There is definitely a male-female difference. All the pops look like they are overlapping when they are all in together. I don't see any major outliers in the PCA. Add a classifier for "subgroup" in the name?