All language subtitles for 67031x_PR_LOD_Scores_05_Unknown_Phases_v2-en

PETER REDDIEN: All right, so let's turn

to phase being unknown and how to deal with this.

OK.

All right, so let's imagine a pedigree here, a simple one.

OK.

And then we have this is our D over plus situation again.

I'll give you the SSR genotypes then.

OK.

OK.

All right, so we've got A over B, A over B, A over

A, B over B, and A over A. Do we have any informative meiosis?

Let's think.

Let's take an individual at a time.

This individual, what did this male offspring

get from this female parent?

It must have gotten A and for the disease gene,

we can tell which went together.

We know that A and D came together from this parent.

That's an informative meiosis.

And you'll see that there are three informative meiosis here.

Do we know the phase of this individual?

No.

We don't have any information about this female's parents.

We have no way of knowing the phase.

OK, so this is a situation where we have data,

but phase is unknown.

What do we do with this data?

Well, we could say, well, we're going

to have a tough time figuring out

recombinant or non-recombinant gametes.

So let's just ignore it.

On the other hand, we want to get every bit of information

we can out of these pedigrees.

It's, kind of, hard to get a lot of data here.

Families tend to be small compared to the type of data

we can get in the lab with model organisms.

And if they kept going a certain direction,

you would intuitively think that's got

to be indicative of something.

It's going to be changing the probability of linkage.

And so we can include these meiosis

in our calculation for the LOD score.

So if the phase is unknown, we're

going to average the probabilities of the data

from the two possible phases.

OK, well, we could say it could have been with A with D,

or it could have been A with plus.

Since we don't know, we'll just say

there's a 50% chance of either scenario.

And then we'll average the probabilities

of getting these data.

OK, so our LOD score then--

OK, just like before, it's going to be now

the probability of data given linkage in phase one

plus the probability of getting the data with linkage

in phase two.

OK, and then we're going to take the average of this

and then divide by the probability of getting

the data if they are unlinked.

OK.

So let's go through that with this new example

where we have three informative meiosis here.

OK, so our LOD score then--

OK, so one scenario I'm going to give you

a theta for this example, just going

to give you a theta of 0.1.

So our data here would be 0.45 cubed plus 0.05 cubed divided

by 0.25 cubed.

OK.

All right, so some phase where they were A with D

would have three non-recombinant gametes,

as you'll see if you walk it through,

and then some phase where it was A with plus

would have three recombinant gametes.

And with theta 0.1, then we'd have this probability

of a given non-recombinant gamete and this probability

of a recombinant gamete, and we'll

see that we get a LOD score here of 0.465.

OK, so you'll see that it is positive.

So these data do contribute it-- do contribute to a LOD score.

They contribute, in this case, a positive value to a LOD score.

What you'll see is that if you have one individual, one

meiosis where the phase is not known like in the example

we had up here if you do this, your LOD score

will be 0 for one meiosis.

So if you just have one individual coming

from one meiosis where the phase was unknown in the parents,

you don't have-- you can ignore it.

But if you have multiple individuals where

the parents had phase unknown like in this scenario,

then they do contribute to the LOD score.