Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
0
00:00:00,000 --> 00:00:02,333
PETER REDDIEN: So we'll have the data of the inheritance
1
00:00:02,333 --> 00:00:03,670
patterns of these repeats.
2
00:00:03,670 --> 00:00:05,430
And we want to know which of these repeats
3
00:00:05,430 --> 00:00:08,670
are linked with the individual-- with the allele
4
00:00:08,670 --> 00:00:10,660
of the gene causing the trait.
5
00:00:10,660 --> 00:00:12,420
So we need some statistical test for that.
6
00:00:12,420 --> 00:00:19,910
And, for that, we're going to use a LOD score, which
7
00:00:19,910 --> 00:00:21,215
is the log of the odds ratio.
8
00:00:21,215 --> 00:00:31,385
9
00:00:31,385 --> 00:00:32,385
So we'll have some data.
10
00:00:32,385 --> 00:00:37,200
11
00:00:37,200 --> 00:00:39,943
And what we want to do with the data
12
00:00:39,943 --> 00:00:41,235
is look at the odds of linkage.
13
00:00:41,235 --> 00:01:02,760
14
00:01:02,760 --> 00:01:06,720
We want to compare the odds of linkage, where we're
15
00:01:06,720 --> 00:01:13,840
given data, versus the odds of whatever we're comparing,
16
00:01:13,840 --> 00:01:16,975
the mark, the SSR in our disease-causing allele,
17
00:01:16,975 --> 00:01:18,100
the odds of being unlinked.
18
00:01:18,100 --> 00:01:30,740
19
00:01:30,740 --> 00:01:34,940
OK, so we can define some events here.
20
00:01:34,940 --> 00:01:47,780
21
00:01:47,780 --> 00:01:51,110
Event X is the marker is linked.
22
00:01:51,110 --> 00:01:54,455
We have not X as unlinked.
23
00:01:54,455 --> 00:01:59,770
24
00:01:59,770 --> 00:02:04,075
And then event Y will be our data in the pedigree.
25
00:02:04,075 --> 00:02:12,450
26
00:02:12,450 --> 00:02:15,810
OK, so we could get an expression that
27
00:02:15,810 --> 00:02:20,310
allows us to get this odds ratio by using Bayes' theorem where
28
00:02:20,310 --> 00:02:22,350
we could say the probability of X
29
00:02:22,350 --> 00:02:31,288
given Y. We could get an odds ratio here,
30
00:02:31,288 --> 00:02:32,830
sort of what I've set up there, where
31
00:02:32,830 --> 00:02:36,250
we get the ratio of the probability of X,
32
00:02:36,250 --> 00:02:40,180
the marker being linked given our data, divided
33
00:02:40,180 --> 00:02:45,135
by the probability of our marker being unlinked, given our data.
34
00:02:45,135 --> 00:02:46,510
That's going to be our odds ratio
35
00:02:46,510 --> 00:02:49,000
that we're trying to calculate.
36
00:02:49,000 --> 00:02:57,170
And so I'm just going to now use Bayes' theorem, which
37
00:02:57,170 --> 00:03:00,568
is listed up there for you, and say this
38
00:03:00,568 --> 00:03:02,360
will be equal to the probability of Y given
39
00:03:02,360 --> 00:03:08,360
X times the probability of X times 1
40
00:03:08,360 --> 00:03:17,550
over the probability of Y divided by the probability of Y
41
00:03:17,550 --> 00:03:23,670
given not X times the probability of not X times 1
42
00:03:23,670 --> 00:03:29,740
over the probability of Y.
43
00:03:29,740 --> 00:03:33,180
This cancels out.
44
00:03:33,180 --> 00:03:45,290
And this is sort of our posterior odds ratio,
45
00:03:45,290 --> 00:03:50,162
the odds of X given Y, given our data.
46
00:03:50,162 --> 00:03:51,245
This is sort of the prior.
47
00:03:51,245 --> 00:03:59,450
48
00:03:59,450 --> 00:04:02,920
And this is a relationship.
49
00:04:02,920 --> 00:04:04,450
What you'll see is that this odds
50
00:04:04,450 --> 00:04:07,120
ratio that we're interested in is directly related
51
00:04:07,120 --> 00:04:08,860
to this ratio.
52
00:04:08,860 --> 00:04:11,470
And this ratio, as you'll see in the next lecture,
53
00:04:11,470 --> 00:04:14,088
is something you can easily determine from the data.
54
00:04:14,088 --> 00:04:15,380
So that's why we're doing this.
55
00:04:15,380 --> 00:04:17,050
We can get the odds ratio we care about
56
00:04:17,050 --> 00:04:20,589
because this is something that we can determine.
57
00:04:20,589 --> 00:04:24,168
This prior is going to be the same, no matter
58
00:04:24,168 --> 00:04:26,210
what your data is or what gene you're looking at.
59
00:04:26,210 --> 00:04:29,200
So we're not really going to consider that further.
60
00:04:29,200 --> 00:04:30,940
But this relationship is what allows
61
00:04:30,940 --> 00:04:36,350
us to get a useful statistical test where our LOD score--
62
00:04:36,350 --> 00:04:39,490
this is the last thing I'm going to write here--
63
00:04:39,490 --> 00:04:49,360
is the log base 10 of the probability of the data,
64
00:04:49,360 --> 00:05:05,360
given linkage, at some distance theta--
65
00:05:05,360 --> 00:05:08,270
OK, and I'm going to go into the details of this equation
66
00:05:08,270 --> 00:05:10,220
in the next lecture--
67
00:05:10,220 --> 00:05:15,755
divided by the probability of getting the data if unlinked.
68
00:05:15,755 --> 00:05:25,710
69
00:05:25,710 --> 00:05:27,090
That is our LOD score.
70
00:05:27,090 --> 00:05:29,740
And that is going to be our statistical test.
71
00:05:29,740 --> 00:05:31,830
So what we're going to do next time,
72
00:05:31,830 --> 00:05:35,070
we're going to have pedigrees.
73
00:05:35,070 --> 00:05:37,230
We'll get data on the SSR genotypes
74
00:05:37,230 --> 00:05:39,125
and the segregation in the pedigree.
75
00:05:39,125 --> 00:05:40,500
And we'll calculate our LOD score
76
00:05:40,500 --> 00:05:43,670
and see if we can find where the gene is.
5422
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.