Skip to content

README.md

Latest commit

 

History

History
105 lines (86 loc) · 3.82 KB

README.md

File metadata and controls

105 lines (86 loc) · 3.82 KB

In this example, we will reanalyze the Arctic Data from Nielsen, Vaughn, et al., PLOS Genetics, 2023.

The input data ArcticData.txt.gz was created by gzipping https://github.com/avaughn271/AdmixtureBayes/blob/main/Data/ArcticData.txt.

Step 1

The first step is to convert the allele frequencies into f2-summary statistics:

../../src/orientagraph \
    -i ArcticData.txt.gz \
    -root UstIshim \
    -f2 \
    -freq2stat \
    -k 5000 \
    -seed 12345 \
    -o arctic-data-k5000-seed12345

which can be given to OrientAGraph as input (note that you do not need to do this step but it can be helpful if you want to explore different versions of running OrientAGraph without re-estimating the f2-statistics).

Note that the author's of TreeMix recommend using a value of -k that far exceeds the known extent of LD in the organism in question.

Before moving on, let's check out the data. Type

gzcat arctic-data-k5000-seed12345.cov.gz

to see the matrix of f2-statistics and type

gzcat arctic-data-k5000-seed12345.covse.gz

to see the matrix of standard errors.

Step 2: Run OrientAGraph in default mode.

Let's run OrientAGraph in default mode.

../../src/orientagraph \
    -i arctic-data-k5000-seed12345.cov.gz \
    -givenmat arctic-data-k5000-seed12345.covse.gz \
    -f2 \
    -root UstIshim \
    -m 4 \
    -seed 12345 \
    -o orientagraph-default-arctic-data-k5000-seed12345 \
    &> orientagraph-default-arctic-data-k5000-seed12345.log

If you look at the log file, you will see

Network search will include 4 admixture edge additions
Exhaustive search will be performed for the following admixture edge additions: 
MLNO search will be performed after each of the following admixture edge additions: 1 2

This means that OrientAGraph will perform MLNO search after the first two edge additions (same as -mlno 1,2). The default changed starting in version 1.2; previously, you needed to use the command -mlno flag to perform any MLNO search. The default was changed because -mlno 1,2 is appropriate for most users. It is may be sufficient to re-route admixture edges away from the outgroup population and it is relatively computationally efficient (even for large numbers of populations or admixture edges).

Step 2: Run OrientAGraph without MLNO search.

Now let's run OrientAGraph without the MLNO search.

../../src/orientagraph \
    -i arctic-data-k5000-seed12345.cov.gz \
    -givenmat arctic-data-k5000-seed12345.covse.gz \
    -f2 \
    -root UstIshim \
    -m 4 \
    -mlno 0 \
    -seed 12345 \
    -o orientagraph-mlno0-arctic-data-k5000-seed12345 \
    &> orientagraph-mlno0-arctic-data-k5000-seed12345.log

Step 3: Visualize the results.

To visualize the residuals, you will need to create a poporder file.

gzcat ArcticData.txt.gz | head -n1 | sed 's/ /\n/g' > poporder.txt

Then, you will need to start R.

The following commands can be used in R

source("../../src/plotting_funcs.R")
plot_tree("orientagraph-default-arctic-data-k5000-seed12345")
plot_resid("orientagraph-default-arctic-data-k5000-seed12345", "poporder.txt")

and repeated for the output without MLNO.

Left image is OrientAGraph in default mode (llik = 364.943); Right image is OrientAGraph witout MLNO (llik = 364.943).

For this dataset, MLNO search does not improve the graph.

To analyze a dataset where MLNO has an impact, go to this example.

If you find real world examples where MLNO makes a difference, we would love to know!! Please send a message to [email protected]