An Intro to Dog DNA Analysis What’s in a...
Transcript of An Intro to Dog DNA Analysis What’s in a...
![Page 1: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/1.jpg)
What’s in a Mutt?An Intro to Dog DNA Analysis
Lecture 4Jan 14th, 2019
![Page 2: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/2.jpg)
RecapOur mutt’s chromosomes are a mosaic, and we’d like to figure out what original purebred dog each piece of DNA came from.
![Page 3: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/3.jpg)
RecapOur mutt’s chromosomes are a mosaic, and we’d like to figure out what original purebred dog each piece of DNA came from.
![Page 4: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/4.jpg)
RecapTo do this we need to phase the SNP data (separate chromosomes).A
AGT
GAAT
![Page 5: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/5.jpg)
Recap1. Where do the chunks begin and end?2. What breed is each chunk?A
AGT
GAAT
![Page 6: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/6.jpg)
Recap: Comparing to purebredsFor now, let’s assume we know what breed each chunk is.
How might we go about determining the
breed of each?
![Page 7: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/7.jpg)
Comparing to purebredsFor now, let’s assume we know what breed each chunk is.
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
![Page 8: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/8.jpg)
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Comparing to purebredsFor now, let’s assume we know what breed each chunk is.
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
Golden and Chow Golden and Shiba Golden and Chow
![Page 9: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/9.jpg)
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Comparing to purebredsFourth combo [ATG] and [GAC] not possible; could be Golden and Unknown
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
Golden and Chow Golden and Shiba Golden and Chow
![Page 10: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/10.jpg)
Comparing to purebreds
How is this picture different from what our purebred data actually look like?
![Page 11: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/11.jpg)
Comparing to purebreds● Six dogs per breed
○ So we see multiple genotypes per purebred
● Phased purebred data○ So we might only see certain allele combinations for adjacent SNPs
ACAT
ACAT
TCAT
ACAA
ACAT
ACAT
ACAT
ACAT
ACAT
TCAT
TCGT
TCGT
![Page 12: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/12.jpg)
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Comparing to purebreds
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
Golden and Chow Golden and Shiba Golden and Chow
Let’s say for SNP3, for goldens, we see G 10% of the time,
shiba: 2%, and chows: 30%.
![Page 13: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/13.jpg)
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Comparing to purebreds
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
Golden and Chow Golden and Shiba Golden and Chow
Let’s say for SNP3, for goldens, we see G 10% of the time,
shiba: 2%, and chows: 30%.
![Page 14: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/14.jpg)
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Comparing to purebreds
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
Golden and Chow Golden and Shiba Golden and Chow
So based on our mutt, the most likely phasing for a golden and a chow with these genotypes is:
Golden: Chow: AAC
GAG
GTG
GTC
![Page 15: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/15.jpg)
Comparing to purebreds● Six dogs per breed
○ So we see multiple genotypes per purebred
● Phased purebred data○ So we might only see certain allele combinations for adjacent SNPs
ACAT
ACAT
TCAT
ACAA
ACAT
ACAT
ACAT
ACAT
ACAT
TCAT
TCGT
TCGT
![Page 16: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/16.jpg)
Comparing to purebreds● Six dogs per breed
● Phased purebred data
ACAT
ACAT
TCAT
ACAA
ACAT
ACAT
ACAT
ACAT
ACAT
TCAT
TCGT
TCGT
Now we have phased purebreds, so we can
use this info too!
![Page 17: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/17.jpg)
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Comparing to purebreds
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
Golden and Chow Golden and Shiba Golden and Chow
Let’s say we only see the following phasing in goldens:
AAC / GAG
![Page 18: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/18.jpg)
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Comparing to purebreds
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
Golden and Chow Golden and Shiba Golden and Chow
Let’s say we only see the following phasing in goldens:
AAC / GAG
![Page 19: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/19.jpg)
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Comparing to purebreds
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
Golden and Chow Golden and Shiba Golden and Chow
Let’s say we only see the following phasing in goldens:
AAC / GAG
Let’s say for SNP3, for goldens, we see G 10% of the time,
shiba: 2%, and chows: 30%.
![Page 20: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/20.jpg)
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
Comparing to purebreds
Compare to haplotypes:
Goldens have AG AA CGShiba Inus have AA TT CCChow chows have GG TT CG
Mutt: AG AT CG
SNP1 SNP2 SNP3
Golden and Chow Golden and Shiba Golden and Chow
Phasing Allele frequencies
![Page 21: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/21.jpg)
Hidden Markov Models (HMMs) with SupportMixWe’ll use a program called SupportMix, which takes in:
1. Phased SNPs from purebred dogs2. Phased SNPs from our mutts3. A “genetic linkage map” of the centiMorgan distances between SNPs
Output: For each mutt, gives the best guess breed for each SNP, and the probability the given guess is correct
Method: Hidden Markov Model
![Page 22: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/22.jpg)
Hidden Markov Models (HMMs) with SupportMixWe’ll use a program called SupportMix, which takes in:
1. Phased SNPs from purebred dogs2. Phased SNPs from our mutts
When we phased our purebred dogs, we also got out mutt phasings. So, we can phase mutts and purebreds together to get phased mutts!
AAC
GTG
Note: We use different sets of purebred dogs to phase the mutts than we use with SupportMix (6 from each breed to phase the mutts, and 6 others from each breed that we phase with each other and/or with other mutts) to compare to.
![Page 23: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/23.jpg)
Hidden Markov Models (HMMs) with SupportMixWe’ll use a program called SupportMix, which takes in:
1. Phased SNPs from purebred dogs2. Phased SNPs from our mutts3. A “genetic linkage map” of the centiMorgan distances between SNPs
Output: For each mutt, gives the best guess breed for each SNP, and the probability the given guess is correct
Method: Hidden Markov Model
![Page 24: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/24.jpg)
Slides adapted from the Jackson Laboratory
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Oversimplified again, let’s consider these the most common haplotype for each breed
![Page 25: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/25.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Slides adapted from the Jackson Laboratory
![Page 26: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/26.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Slides adapted from the Jackson Laboratory
![Page 27: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/27.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Slides adapted from the Jackson Laboratory
![Page 28: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/28.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Slides adapted from the Jackson Laboratory
![Page 29: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/29.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Slides adapted from the Jackson Laboratory
![Page 30: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/30.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Slides adapted from the Jackson Laboratory
![Page 31: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/31.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Slides adapted from the Jackson Laboratory
![Page 32: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/32.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Slides adapted from the Jackson Laboratory
![Page 33: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/33.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C AA G T G G C G T A TT C T G T C G A C T
A C G T T C G A C T
BeagleCollie
Poodle
Fido
HMMs
Represent ancestry by painting with the breed color
One chromosome
Slides adapted from the Jackson Laboratory
![Page 34: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/34.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C A.82 .85 .74 .95 .90 .89 .81 .75 .91 .94
T C T G T C G A C T.58 .91 .93 .79 .84 .85 .92 .78 .86 .99
A C G T T C G A C T
Beagle
Poodle
Fido
HMMs
Represent ancestry by painting with the breed color
One chromosome
Slides adapted from the Jackson Laboratory
![Page 35: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/35.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C A.82 .85 .74 .95 .90 .89 .81 .75 .91 .94
T C T G T C G A C T.58 .91 .93 .79 .84 .85 .92 .78 .86 .99
A C G T T C G A C T
Beagle
Poodle
Fido
HMMs
Represent ancestry by painting with the breed color
One chromosome
Slides adapted from the Jackson Laboratory
![Page 36: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/36.jpg)
SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9 SNP10
A C G T T C G T C A.82 .85 .74 .95 .90 .89 .81 .75 .91 .94
T C T G T C G A C T.58 .91 .93 .79 .84 .85 .92 .78 .86 .99
A C G T T C G A C T
Beagle
Poodle
Fido
HMMs
Represent ancestry by painting with the breed color
One chromosome
It seems unlikely we’d transition for one SNP and then transition back. HMMs account for this!
Slides adapted from the Jackson Laboratory
![Page 37: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/37.jpg)
● Goal: Determine the most probable path through the data.
○ Translation: Determine the most probable breed along each haplotype. Maximize Pr(breed|data)
https://onlinecourses.science.psu.edu/stat857/node/203
Markers
Boxer
Collie
Poodle
Vizsla
HMM: Viterbi Decoding
Slides adapted from the Jackson Laboratory
![Page 38: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/38.jpg)
https://onlinecourses.science.psu.edu/stat857/node/203
Boxer
Collie
Poodle
Vizsla
● To determine the most probable path, we take into account probabilities of seeing a SNP given a breed, but we also consider the probability of transitioning breed.
HMM: Viterbi Decoding
Slides adapted from the Jackson Laboratory
![Page 39: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/39.jpg)
HMMs deal with data, which we call emissions, and hidden states, which is what we’re trying to determine.
Emissions: SNPsHidden States: Breeds
A T T G C G A A
Hidden Markov Models (HMMs)
![Page 40: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/40.jpg)
A T T G C G A A
Hidden Markov Models (HMMs)
? ? ? ? ? ? ? ?93
![Page 41: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/41.jpg)
A T T G C G A A
Hidden Markov Models (HMMs)
? ? ? ? ? ? ? ?93
How likely is it I see “A” if the hidden state is a … husky? corgi? chow? Etc.
![Page 42: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/42.jpg)
A T T G C G A A
Hidden Markov Models (HMMs)
? ? ? ? ? ? ? ?93
How likely is it I see “A” if the hidden state is a … husky? corgi? chow? Etc.
Emission probabilities: P(A1|husky)P(A1|corgi)...
P(T2|husky)P(T2|corgi)...
1 2 3 4 5 6 7 8
P(allelen|husky)P(allelen|corgi)...
...
![Page 43: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/43.jpg)
A T T G C G A A
Hidden Markov Models (HMMs)
? ? ? ? ? ? ? ?93
If the current breed is husky, how likely is it the breed at the next SNP site is … husky? corgi? chow? etc
1 2 3 4 5 6 7 8
![Page 44: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/44.jpg)
A T T G C G A A
Hidden Markov Models (HMMs)
? ? ? ? ? ? ? ?93
If the current breed is husky, how likely is it the breed at the next SNP site is … husky? corgi? chow? Etc
Transition probabilities: Because we know we have linked regions inherited together, intuitively P(huskyi|huskyi-1) > P(corgii|huskyi-1)
1 2 3 4 5 6 7 8
![Page 45: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/45.jpg)
Hidden Markov Models (HMMs)How do we get transition probabilities?
Based on what we know, we can intuit that:
1. Probability breed_A --> breed_B is the same regardless of breed (A != B)
2. It seems like it’s a higher probability that breed_A --> breed_A.
So we don’t need transition probabilities for all breeds --> all breeds!
![Page 46: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/46.jpg)
Hidden Markov Models (HMMs)How do we get transition probabilities?
We know two SNPs are more likely to be in the same “chunk” if they are nearby one another. We have centiMorgan distances between all our SNPs.
![Page 47: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/47.jpg)
Hidden Markov Models (HMMs)How do we get transition probabilities?
1. Probability breed_A --> breed_B is the same regardless of breed (A != B)
2. It seems like it’s a higher probability that breed_A --> breed_A.
3. We know two SNPs are more likely to be in the same “chunk” if they are nearby one another. We have centiMorgan distances between all our SNPs.
We can calculate probabilities from this!
![Page 48: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/48.jpg)
Hidden Markov Models (HMMs)How do we get transition probabilities? Another way would to train the HMM on a labeled mutt.
If we have a mutt and we know what it’s ancestral segments are, we can examine that data to determine how likely breed transitions are to occur at different cM distances.
A T T G C G A A
![Page 49: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/49.jpg)
https://onlinecourses.science.psu.edu/stat857/node/203
Boxer
Collie
Poodle
Vizsla
1. Examine all possible hidden state paths (breed assignments)2. Use emission and transition probabilities to choose the path that
maximizes the probability of the entire sequence (Viterbi)
HMM: Viterbi Decoding
Slides adapted from the Jackson Laboratory
![Page 50: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/50.jpg)
Final HMM NotesThe way we calculate using the probabilities assumes that the state (breed) at a given SNP is only dependent on the state (breed) of the SNP before it.
HMMs are used for a lot of other biology applications, including gene finding in bacteria.
To learn about them in more detail (and code your own!), take Computational Genomics (EN 601.439/639) with Ben Langmead in Fall 2019!
![Page 51: An Intro to Dog DNA Analysis What’s in a Mutt?ccb.jhu.edu/people/rsherman/EN.601.147.13_Lecture_4.pdfAn Intro to Dog DNA Analysis Lecture 4 Jan 14th, 2019 Recap Our mutt’s chromosomes](https://reader034.fdocuments.in/reader034/viewer/2022052022/603730c0607eb2299f43b915/html5/thumbnails/51.jpg)
Project LogisticsToday: More data exploration (continue part 1 and/or part 2)Wed/Fri: Finding Clarence, Reilly, and Finch’s breedsNext week: Concept exploration (no coding, but you’ll need laptops)
Part 1 due Wednesday, Jan 16.Part 2 due Friday, Jan 18.
Please turn in your code and question answers to [email protected] and include EN.601.147 in the subject line.
Make sure both your names are on your writeups!