Tutorial for autosomal calculators

I ll make this quick in 5 simple steps, since so many people ask how to use these calculators out of GEDmatch


1- First, Download R and Dodecad DIY:

http://www.r-project.org/

https://drive.google.com/file/d/0B7A...4NjRmMTlk/view


2- Get your genome raw datas ready in your Dodecad folder in the form of :

Genome_Yourname.zip
Genome_Yourname.txt (unzipped)

Drag and drop the calculators files inside this folder as well:

Spoiler!

Calculator files:
https://docs.google.com/file/d/0B7AJ...hl=en_US&pli=1

3- Open R and "change current folder", point to the calculator's folder

4- Since it's the first time you are using it, you gonna have to standardize your data :

Quote:

---

Enter the command : source('standardize.r')

This loads a small program that will convert your data from the
company-specific format to a common format in the next step.

- At the R prompt, enter:

a. If you have 23andMe data (either v2 or v3 chip):

standardize('johndoe.txt', company='23andMe')

b. if you have Family Finder data (Illumina chip only):

standardize('johndoe.csv', company='ftdna')

This command will write a file called 'genotype.txt' in the working directory;
this contains your genotype in a format understood by DIYDodecad.

---

5- Finally you run the calculator with this command :

system('DIYDodecadWin dv3.par')

Replace the "dv3" part by the name of the calculator you are currently using

For example with Dodecad "Euro7" :

system('DIYDodecadWin Euro7.par')

The program is gonna stall for a minute or two depending on the calculator (means it's working) and then you get your admix results which might look like this (not as cool of course) :

FINAL ADMIXTURE PROPORTIONS:
----------------------------

3.03% Caucasus
44.43% Northwestern
17.76% Northeastern
14.39% Southeastern
0.00% African
0.02% Far_Asian
20.36% Southwestern

CPU time = 69.56 sec



How do you Oracle then ?

Download Admix4 : https://drive.google.com/file/d/0B9o...RWTU1GdEE/view

Oracle4 file : Attachment 59268

Put the Oracle calculator file in the folder, everytime you will launch Admix4, it s gonna ask which calculator file you want to use. Once the right one is loaded, just input you final admix results, let it process it, you ll get a file named "Output1" in that same folder with your single and multi population results.

good explained and detailed :thumbs up this thread need a sticky, would helpfull for the other testers

Stickied.

Bump! Good tutorial! :thumb001:

ANOTHER TUTORIAL:

http://www.theapricity.com/forum/sho...=1#post4330097

Nope, it is Steppe K10. This calculator is not on GEDmatch.

But you can use it on your computer, following these tips:

HOW TO?:

Let’s presume that John tested with 23andMe, and that Adam tested with FTDNA.

John will download his 23andMe raw data and name the file as John.txt , while Adam will download his FTDNA raw data (Build 37 Raw Data Concatenated) and name the file as Adam.csv . Both using Windows.

Download SteppeK10.zip (save icon - an arrow in top right):
https://drive.google.com/file/d/0B-X...lRSkcwV0k/view

Extract the files into a folder called SteppeK10, in C:/ . Now you have the folder C:/SteppeK10 , right? Inside it, the files DIYDodecadWin.exe , standardize.r , Steppe.10.F , Steppe.alleles , steppe.par , Steppe.txt .

John and Adam will put the files John.txt and Adam.csv into the folder C:/SteppeK10, respectively.

Download R software: https://cran.r-project.org/bin/windows/base/

Install* this program in English and using default options (if you don’t know how to do it, just ask). After that, run it (click in the icon “R x.. 3.3.1” in Desktop).
*I installed just the 64 bits version.

Once R is running, click in “File” (top left), then click in “Change dir”. Select the folder C:/SteppeK10 and click in “ok”.

Do you see the R console? After the “>” character, type the command below and press Enter:
source('standardize.r')

Now:
a. John, who tested hypothetically with 23andMe, will type the command below, press Enter and wait til he can type anything again:
standardize('John.txt', company='23andMe')

b. Adam will do that with this command:
standardize('Adam.csv', company='ftdna')

Finally, you both will type the command below and wait results show up:
system('DIYDodecadWin steppe.par')

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

Spreadsheet with population averages:

https://docs.google.com/spreadsheets...gid=1809893991

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

Official Tutorial:

http://dodecad.blogspot.com/2011/09/...ecad-v-21.html

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

Turkic K11 calculator:

http://www.theapricity.com/forum/sho...=1#post4068246

I don't understand...

Here you can download Eurogenes K36 DIY calculator files:

https://onedrive.live.com/?authkey=%...&action=locate

Someone should make a Youtube tutorial video :)

Is it possible to run gedmatch kit numbers in calculators outside gedmatch?

Quote:

---

Originally Posted by Litvin

ANOTHER TUTORIAL:

http://www.theapricity.com/forum/sho...=1#post4330097

Nope, it is Steppe K10. This calculator is not on GEDmatch.

But you can use it on your computer, following these tips:

HOW TO?:

Let’s presume that John tested with 23andMe, and that Adam tested with FTDNA.

John will download his 23andMe raw data and name the file as John.txt , while Adam will download his FTDNA raw data (Build 37 Raw Data Concatenated) and name the file as Adam.csv . Both using Windows.

Download SteppeK10.zip (save icon - an arrow in top right):
https://drive.google.com/file/d/0B-X...lRSkcwV0k/view

Extract the files into a folder called SteppeK10, in C:/ . Now you have the folder C:/SteppeK10 , right? Inside it, the files DIYDodecadWin.exe , standardize.r , Steppe.10.F , Steppe.alleles , steppe.par , Steppe.txt .

John and Adam will put the files John.txt and Adam.csv into the folder C:/SteppeK10, respectively.

Download R software: https://cran.r-project.org/bin/windows/base/

Install* this program in English and using default options (if you don’t know how to do it, just ask). After that, run it (click in the icon “R x.. 3.3.1” in Desktop).
*I installed just the 64 bits version.

Once R is running, click in “File” (top left), then click in “Change dir”. Select the folder C:/SteppeK10 and click in “ok”.

Do you see the R console? After the “>” character, type the command below and press Enter:
source('standardize.r')

Now:
a. John, who tested hypothetically with 23andMe, will type the command below, press Enter and wait til he can type anything again:
standardize('John.txt', company='23andMe')

b. Adam will do that with this command:
standardize('Adam.csv', company='ftdna')

Finally, you both will type the command below and wait results show up:
system('DIYDodecadWin steppe.par')

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

Spreadsheet with population averages:

https://docs.google.com/spreadsheets...gid=1809893991

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

Official Tutorial:

http://dodecad.blogspot.com/2011/09/...ecad-v-21.html

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

Turkic K11 calculator:

http://www.theapricity.com/forum/sho...=1#post4068246

---

error message after entering the following

standardize('f.txt', company='23andMe')

ERROR: could not find function "standardize"



Standardize.r is in the folder? What am I doing wrong? K11 folder is in C as instructed?

Quote:

---

Originally Posted by Zuvaq

Is it possible to run gedmatch kit numbers in calculators outside gedmatch?

---

No, you need the raw datas, gedmatch numbers are just numbers tied to people's genomes in their database.

Quote:

---

Originally Posted by Dibran

error message after entering the following

standardize('f.txt', company='23andMe')

ERROR: could not find function "standardize"



Standardize.r is in the folder? What am I doing wrong? K11 folder is in C as instructed??

---

I don't know if it's your error but you should run source('standardize.r') first and getting a new prompt, then standardize('f.txt', company='23andMe')


While f.txt has to be replaced by the actual name of your genome like Dibran16022017.txt

Quote:

---

Originally Posted by Zuvaq

Is it possible to run gedmatch kit numbers in calculators outside gedmatch?

---

well, this thread actually does not exist :rolleyes:

does it work with raw data from Ancestry?

Very nice explanation, thank you very much!

im trying to do this for a new calculator and im getting really mad because I cant get it to work, I have followed all the steps and I get no damn results

These calculators are boring. The only thjng Im still interested is about haplogroups for genetics

+

How to make 3D PCA in RStudio and perform hierarchical k-means clustering
https://youtu.be/_lZ_EqV-cZw
https://youtu.be/_lZ_EqV-cZw

Eurogenes K13 datasheet: https://drive.google.com/file/d/1cDv...fwoAHewJ6/view

Script:

Spoiler!

Code:

---

install.packages("randomcoloR")
install.packages("factoextra")
install.packages("rgl")
install.packages("rmarkdown", repos = "https://cran.revolutionanalytics.com")

data01 <- read.csv(file.choose(), row.names = 'Population')
pca <- princomp(data01[,1:13], cor=TRUE, scores=TRUE)
summary(pca)
library(rgl) #load the package
plot3d(pca$scores[,1:3], col="red", type="p")
text3d(pca$scores[,1:3],texts=rownames(data01),font=2)
grid3d('x')
grid3d('y')
grid3d('z')

text3d(pca$loadings[,1:3], texts=rownames(pca$loadings), col="red")
coords <- NULL
for (i in 1:nrow(pca$loadings)) {
  coords <- rbind(coords, rbind(c(0,0,0),pca$loadings[i,1:3]))
}
lines3d(coords, col="red", lwd=4)

library(factoextra) #load the package
cl <- hkmeans(data01[,1:13],130) #130 clusters
data01$cluster <- as.factor(cl$cluster)
library(randomcoloR) #load the package
palette <- distinctColorPalette(130) #130 colors
palette(palette)
plot3d(pca$scores[,1:3], col=data01$cluster, main="Hierarchical k-means clusters")
text3d(pca$scores[,1:3],texts=rownames(data01),font=2, col=data01$cluster)

observer3d(0, 0, 40) # x, y, z

---

__________________________________________

and

How to make custom oracle in RStudio (Gedmatch, G25...)

https://www.youtube.com/watch?v=cFpBinU5E18
https://www.youtube.com/watch?v=cFpBinU5E18

Algorithm: https://docs.google.com/document/d/1...gAZMCzx8A/edit

Quote:

---

Originally Posted by Lemgrant

+

How to make 3D PCA in RStudio and perform hierarchical k-means clustering
https://youtu.be/_lZ_EqV-cZw
https://youtu.be/_lZ_EqV-cZw

Eurogenes K13 datasheet: https://drive.google.com/file/d/1cDv...fwoAHewJ6/view

Script:

Spoiler!

Code:

---

install.packages("randomcoloR")
install.packages("factoextra")
install.packages("rgl")
install.packages("rmarkdown", repos = "https://cran.revolutionanalytics.com")

data01 <- read.csv(file.choose(), row.names = 'Population')
pca <- princomp(data01[,1:13], cor=TRUE, scores=TRUE)
summary(pca)
library(rgl) #load the package
plot3d(pca$scores[,1:3], col="red", type="p")
text3d(pca$scores[,1:3],texts=rownames(data01),font=2)
grid3d('x')
grid3d('y')
grid3d('z')

text3d(pca$loadings[,1:3], texts=rownames(pca$loadings), col="red")
coords <- NULL
for (i in 1:nrow(pca$loadings)) {
  coords <- rbind(coords, rbind(c(0,0,0),pca$loadings[i,1:3]))
}
lines3d(coords, col="red", lwd=4)

library(factoextra) #load the package
cl <- hkmeans(data01[,1:13],130) #130 clusters
data01$cluster <- as.factor(cl$cluster)
library(randomcoloR) #load the package
palette <- distinctColorPalette(130) #130 colors
palette(palette)
plot3d(pca$scores[,1:3], col=data01$cluster, main="Hierarchical k-means clusters")
text3d(pca$scores[,1:3],texts=rownames(data01),font=2, col=data01$cluster)

observer3d(0, 0, 40) # x, y, z

---

__________________________________________

and

How to make custom oracle in RStudio (Gedmatch, G25...)

https://www.youtube.com/watch?v=cFpBinU5E18
https://www.youtube.com/watch?v=cFpBinU5E18

Algorithm: https://docs.google.com/document/d/1...gAZMCzx8A/edit

---

Did you know that MyTrueAncestry used your tutorial to make their 3d plot lol?:)

Quote:

---

Originally Posted by Lukasz

Did you know that MyTrueAncestry used your tutorial to make their 3d plot lol?:)

---

They have 3d plots? I saw only 2d pca. I made this so that everyone could make a 3d pca, since RStudio is a free software.
https://www.rstudio.com/products/rstudio/download/

Quote:

---

Originally Posted by Lemgrant

+

How to make 3D PCA in RStudio and perform hierarchical k-means clustering
https://youtu.be/_lZ_EqV-cZw
https://youtu.be/_lZ_EqV-cZw

Eurogenes K13 datasheet: https://drive.google.com/file/d/1cDv...fwoAHewJ6/view

Script:

Spoiler!

Code:

---

install.packages("randomcoloR")
install.packages("factoextra")
install.packages("rgl")
install.packages("rmarkdown", repos = "https://cran.revolutionanalytics.com")

data01 <- read.csv(file.choose(), row.names = 'Population')
pca <- princomp(data01[,1:13], cor=TRUE, scores=TRUE)
summary(pca)
library(rgl) #load the package
plot3d(pca$scores[,1:3], col="red", type="p")
text3d(pca$scores[,1:3],texts=rownames(data01),font=2)
grid3d('x')
grid3d('y')
grid3d('z')

text3d(pca$loadings[,1:3], texts=rownames(pca$loadings), col="red")
coords <- NULL
for (i in 1:nrow(pca$loadings)) {
  coords <- rbind(coords, rbind(c(0,0,0),pca$loadings[i,1:3]))
}
lines3d(coords, col="red", lwd=4)

library(factoextra) #load the package
cl <- hkmeans(data01[,1:13],130) #130 clusters
data01$cluster <- as.factor(cl$cluster)
library(randomcoloR) #load the package
palette <- distinctColorPalette(130) #130 colors
palette(palette)
plot3d(pca$scores[,1:3], col=data01$cluster, main="Hierarchical k-means clusters")
text3d(pca$scores[,1:3],texts=rownames(data01),font=2, col=data01$cluster)

observer3d(0, 0, 40) # x, y, z

---

__________________________________________

and

How to make custom oracle in RStudio (Gedmatch, G25...)

https://www.youtube.com/watch?v=cFpBinU5E18
https://www.youtube.com/watch?v=cFpBinU5E18

Algorithm: https://docs.google.com/document/d/1...gAZMCzx8A/edit

---

thanks so much .... btw I like giordano bruno too c;

I never try R or any data analysis before.... I didn't have any project on mind ... until now ....

I'm autodidact but can you ask you about ... nmonte or how I can create my costume G25 from raw files ?

I mean I just need some inside about how to translate data into coordinates ... I saw in other post about normalize data...

I want to made coordinates of my chromosomes... where can I start to do so ? Dx