Week2 at Vrije Universiteit Amsterdam | Flashcards & Summaries

Lernmaterialien für Week2 an der Vrije Universiteit Amsterdam

Greife auf kostenlose Karteikarten, Zusammenfassungen, Übungsaufgaben und Altklausuren für deinen Week2 Kurs an der Vrije Universiteit Amsterdam zu.

TESTE DEIN WISSEN

Replication Outcome: Association to 

same trait, 

same gene, 

different SNP (or haplotypes)

Lösung anzeigen
TESTE DEIN WISSEN

"Allelic heterogeneity"


Lösung ausblenden
TESTE DEIN WISSEN

SMARTPCA

Lösung anzeigen
TESTE DEIN WISSEN

software for identification of ancestral outliers

(population stratification)



Lösung ausblenden
TESTE DEIN WISSEN

What is PLINK used for?

Lösung anzeigen
TESTE DEIN WISSEN

Quality control of data in GWAS

to reduce possibility to get

1. false-positive

2. false-negative

associations


Lösung ausblenden
TESTE DEIN WISSEN

What is the step before the Quality Control protocol?

 

Lösung anzeigen
TESTE DEIN WISSEN

genotype calling (= genotyping = genotype determination)




Lösung ausblenden
TESTE DEIN WISSEN

What is genotype calling, what is done in this step of the GWAS analysis?

Lösung anzeigen
TESTE DEIN WISSEN

= determination of genotypes

steps:

1. calculate probabilities that a genotype is f.i. aa/aA/AA
(calculate with Illuminus/Chiamo)

2. set threshold

3. genotypes above this threshold are accepted (="called")

4. genotypes below are not accepted (="uncalled")

5. manual inspection of cluster plots



Lösung ausblenden
TESTE DEIN WISSEN

What happens if the threshold in genotyping is set too high / too low?

Lösung anzeigen
TESTE DEIN WISSEN

threshold too high

  => noise


threshold too low

  => "informative missingness" 

(= failure to call may be due to genotype)

  => reduction of power

(many unnecessarily uncalled ("missing") genotypes


Lösung ausblenden
TESTE DEIN WISSEN

Levels of Quality Control?

Lösung anzeigen
TESTE DEIN WISSEN

1. do QC on 'per-individual' level
               & remove inappropriate individuals
2. do QC on 'per-marker' level
               & remove inappropriate SNP's



conservative alternative:
1. QC per-individual & per-marker
        & remove individuals, markers
BUt data may be removed unnecessarily like this!

Lösung ausblenden
TESTE DEIN WISSEN

Per-individual Quality Control:

4 identification steps

Lösung anzeigen
TESTE DEIN WISSEN

Per Individual QC:

1. non-harmonic sex information


2. outlying missing genotype

or heterozygosity rates

3. duplicated or related individuals


4. divergent ancestry 


Lösung ausblenden
TESTE DEIN WISSEN

Per-individual Quality Control

Step 1

non-harmonic sex information

What is done?

Lösung anzeigen
TESTE DEIN WISSEN

correct for wrongly assigned sex.

look at the hemizygosity rate

for males =~ 1

for females =~ 0 


in males, heterozygous markers outside the 

pseudoautosomal region on 23 chromosome are 

marked as "missing" 

(because they should be hemizygous in this regions, having only one X chromosome)

=> watch out for females with many "missing" and correct if possible / exclude if not

Lösung ausblenden
TESTE DEIN WISSEN

Per-individual Quality Control: 

Step 2

missing genotypes or heteroszygosity rates

Lösung anzeigen
TESTE DEIN WISSEN

1. outlying missing genotype:

because DNA was not fully sequenced/DNA sample was contaminated


 2. or heterozygosity rates

(N-O)/N 

N.... called-

O... observed homozygous genotypes


in how many SNP loci are the SNPs not the same/the same

Lösung ausblenden
TESTE DEIN WISSEN

Per-individual Quality Control:

Step 3

duplicated or related individuals

Lösung anzeigen
TESTE DEIN WISSEN

remove duplicated/related individuals*:


1. create "identity by descent" matrix (=IDB; average proportion of shared independent SNPs; scores for all individuals)

2. identify pairs of individuals with IBD > 0.1875

3. keep only the one of the related with IBD > 0.1875

(the one with the most SNPs is usually kept)



IDB scores:

IDB = 1     (monozygotic twins/duplicates)

IDB = 0.5 (dizygotic twins/siblings)

IDB = 0.25 (2nd degree relatives)

IDB = 0.125 (3rd degree relatives)


*only applicable when LD is taken into account

Lösung ausblenden
TESTE DEIN WISSEN

High missingness of SNP's in one individual could be due to ... ?

Lösung anzeigen
TESTE DEIN WISSEN

-  DNA quality is bad


Lösung ausblenden
  • 9742 Karteikarten
  • 119 Studierende
  • 2 Lernmaterialien

Beispielhafte Karteikarten für deinen Week2 Kurs an der Vrije Universiteit Amsterdam - von Kommilitonen auf StudySmarter erstellt!

Q:

Replication Outcome: Association to 

same trait, 

same gene, 

different SNP (or haplotypes)

A:

"Allelic heterogeneity"


Q:

SMARTPCA

A:

software for identification of ancestral outliers

(population stratification)



Q:

What is PLINK used for?

A:

Quality control of data in GWAS

to reduce possibility to get

1. false-positive

2. false-negative

associations


Q:

What is the step before the Quality Control protocol?

 

A:

genotype calling (= genotyping = genotype determination)




Q:

What is genotype calling, what is done in this step of the GWAS analysis?

A:

= determination of genotypes

steps:

1. calculate probabilities that a genotype is f.i. aa/aA/AA
(calculate with Illuminus/Chiamo)

2. set threshold

3. genotypes above this threshold are accepted (="called")

4. genotypes below are not accepted (="uncalled")

5. manual inspection of cluster plots



Mehr Karteikarten anzeigen
Q:

What happens if the threshold in genotyping is set too high / too low?

A:

threshold too high

  => noise


threshold too low

  => "informative missingness" 

(= failure to call may be due to genotype)

  => reduction of power

(many unnecessarily uncalled ("missing") genotypes


Q:

Levels of Quality Control?

A:

1. do QC on 'per-individual' level
               & remove inappropriate individuals
2. do QC on 'per-marker' level
               & remove inappropriate SNP's



conservative alternative:
1. QC per-individual & per-marker
        & remove individuals, markers
BUt data may be removed unnecessarily like this!

Q:

Per-individual Quality Control:

4 identification steps

A:

Per Individual QC:

1. non-harmonic sex information


2. outlying missing genotype

or heterozygosity rates

3. duplicated or related individuals


4. divergent ancestry 


Q:

Per-individual Quality Control

Step 1

non-harmonic sex information

What is done?

A:

correct for wrongly assigned sex.

look at the hemizygosity rate

for males =~ 1

for females =~ 0 


in males, heterozygous markers outside the 

pseudoautosomal region on 23 chromosome are 

marked as "missing" 

(because they should be hemizygous in this regions, having only one X chromosome)

=> watch out for females with many "missing" and correct if possible / exclude if not

Q:

Per-individual Quality Control: 

Step 2

missing genotypes or heteroszygosity rates

A:

1. outlying missing genotype:

because DNA was not fully sequenced/DNA sample was contaminated


 2. or heterozygosity rates

(N-O)/N 

N.... called-

O... observed homozygous genotypes


in how many SNP loci are the SNPs not the same/the same

Q:

Per-individual Quality Control:

Step 3

duplicated or related individuals

A:

remove duplicated/related individuals*:


1. create "identity by descent" matrix (=IDB; average proportion of shared independent SNPs; scores for all individuals)

2. identify pairs of individuals with IBD > 0.1875

3. keep only the one of the related with IBD > 0.1875

(the one with the most SNPs is usually kept)



IDB scores:

IDB = 1     (monozygotic twins/duplicates)

IDB = 0.5 (dizygotic twins/siblings)

IDB = 0.25 (2nd degree relatives)

IDB = 0.125 (3rd degree relatives)


*only applicable when LD is taken into account

Q:

High missingness of SNP's in one individual could be due to ... ?

A:

-  DNA quality is bad


Week2

Erstelle und finde Lernmaterialien auf StudySmarter.

Greife kostenlos auf tausende geteilte Karteikarten, Zusammenfassungen, Altklausuren und mehr zu.

Jetzt loslegen

Das sind die beliebtesten Week2 Kurse im gesamten StudySmarter Universum

Week 2

Griffith University

Zum Kurs
Week 3

Leiden University

Zum Kurs
Week 4

Leiden University

Zum Kurs
Week 1

ZHAW - Zürcher Hochschule für Angewandte Wissenschaften

Zum Kurs
Week4

Vrije Universiteit Amsterdam

Zum Kurs

Die all-in-one Lernapp für Studierende

Greife auf Millionen geteilter Lernmaterialien der StudySmarter Community zu
Kostenlos anmelden Week2
Erstelle Karteikarten und Zusammenfassungen mit den StudySmarter Tools
Kostenlos loslegen Week2