Simple examples in Founders.rmd

[RosCraddock]

Based on the same pedigree, provided simple examples considered in the 24/10 meeting as requested in #7. Additional text towards drafting the founders vignette, but still a work in progress.

[gregorgorjanc]

@RosCraddock will have a look. See what I have done in 2f49d7f regarding demonstrating how partitioning works - I know you are planning to do something similar for the metafounders.

[gregorgorjanc]

The devel version now has a very nicely worked out example in the introduction (see

AlphaPart/vignettes/intro.Rmd

Line 91 in 25d4937

\begin{align*}

) and I have also included a new unit test based on this example (see

AlphaPart/tests/testthat/test-alphapart.R

Line 652 in 25d4937

test_that("Test computation - 2nd example", {

).

I think it would be instructive to do similar for the UPG/metafounder example(s) you are working with. Perhaps you don't do all the examples at this level of detail.

[RosCraddock]

The devel version now has a very nicely worked out example in the introduction (see

AlphaPart/vignettes/intro.Rmd

Line 91 in 25d4937

\begin{align*}

) and I have also included a new unit test based on this example (see

AlphaPart/tests/testthat/test-alphapart.R

Line 652 in 25d4937

test_that("Test computation - 2nd example", {

).
I think it would be instructive to do similar for the UPG/metafounder example(s) you are working with. Perhaps you don't do all the examples at this level of detail.

I agree! I intend to walk through one of the examples to demonstrate the theory with including unknown parent groups. I'll refer to the other vignettes to keep consistency.

[gregorgorjanc]

I agree! I intend to walk through one of the examples to demonstrate the theory with including unknown parent groups. I'll refer to the other vignettes to keep consistency.

I wonder if we need to do it just for one example (for 1 OR 2+ metafounders) or for two (for 1 AND 2+ metafounders)?

I am thinking about this to make it clear how we deal non-zero mean(s) of the base population when there is 1 or 2+ metafounders. Need to implement that table we have developed on the wall to clarify my mind about this;)

[RosCraddock]

I wonder if we need to do it just for one example (for 1 OR 2+ metafounders) or for two (for 1 AND 2+ metafounders)?

I am thinking about this to make it clear how we deal non-zero mean(s) of the base population when there is 1 or 2+ metafounders. Need to implement that table we have developed on the wall to clarify my mind about this;)

I just added the theory to the Founders.rmd. I followed your example from the intro for one unknown parent group and provide similar context for multiple unknown parent groups (but not as detailed).

[RosCraddock]

Hi @gregorgorjanc,
Implemented some changes following Friday's meeting. To summarise:

AlphaPart.R

• New parameter: UPGname with default as "UPG". This gives users flexibility as to how the unknown parent groups/metafounders are named.
• Error handling of unknown parent groups/metafounders, considering whether each UPG have their own record, own path, both parents unknown, etc.

Founder.rmd

• Changes to inform users of the UPGname
• Clarification on what the UPG trends represent

References.rmd

• Two references linked to UPGs.

Next step:
Formal testing. Completed testing locally, but need to update and integrate into the testthat. To do so, will also review and update other tests, as some now issue a warning to consider centring or using UPGs.