Chapter 3 - Exome Sequencing

Question 1

Limiting factors of traditional gene-discovery strategies (linkage mapping and cadidate gene resequencing)

Answer 1

-Availability of small number of cases
-Reduced penetrance
-Locus heterogeneity
-Substantially diminished reproductive fitness
-Responsible mutation may be de novo

Question 2

Mendelian disorders

Answer 2

Inherited disorders like cystic fibrosis (kinkhoest), sickle cell anaemia

Question 3

Coding variation analysis > massively parallel DNA sequencing >

Answer 3

Exome sequencing

Question 4

Limitation of exome sequencing

Answer 4

it does not assess the impact of the non-coding alleles, but discovery of rare alleles underlying Mendelian phenotypes and complex traits

Question 5

Why is exome sequencing effective for detecting rare alleles in Mendelian disorders?

Answer 5

Positional cloning studies are succesful for monogenic disorders
> most alleles underlying Mendelian disorders are protein coding
> large fraction of the rare protein altering variants are predicted to have functional consequences
> splice acceptor and donor sites are enriched for highly functional variation (targeted in exome sequencing)

Question 6

How is the exome defined?

Answer 6

By the entire RefSeq and a large number of hypothetical proteins (this has limitations)

Question 7

Limitations exome defining

Answer 7

-incomplete overview of protein-coding exons
-variety in efficiency of capture probes
-not all templates are sequences efficiently
-not all sequences can be uniquely aligned to the reference genome

Question 8

Wet-lab workflow for exome sequencing

Answer 8

1. Genomic DNA is sheared and used for in vitro shotgun library
2. library fragments are flanked by adapters
3. enrichment for sequences corresponding to exons > aqueous-phase hybridized capture
4. recovery of hybridized fragments by biotin-streptavidin pulldown and washing
5. amplification and massively parallel sequencing
6. Mapping > calling of candidate causal variants

Question 9

Bioinformatics steps in exome sequencing

Answer 9

1. Probe design
2. Quality control
3. Map reads
4. Determine variants
5. Annotate variants
6. Filter known variants
7. exome comparison
8. validation of candidate genes

Question 10

Probe design

Answer 10

Designing probes for capturing exon fragments > unique and efficient probes

Question 11

Quality control

Answer 11

High base quality and equal nucleotide frequencies across the sequence

Question 12

Mapping the reads (bwa)

Answer 12

mapping against reference genome by algorithm
> unmapped reads are discarded, non-unique as well. Low confidence reads may cause problems

Question 13

Determine variants (varscan)

Answer 13

Difference detection compared to reference genome: potential variant or sequencing error.

Question 14

Criteria varscan

Answer 14

1. At position of the variant at least N reads (default 8)
2. From the N reads at least K reads with variant (default 2)
3. Average base quality at position of the variant at least Q (default 15)

Question 15

Annotate variants

Answer 15

Each variant is assigned various properties; gene name, region, nucleotide position, type of mutation, number of reads, quality etc.

Question 16

Filter the known variants

Answer 16

Remove synonymous variants and variants which are present in public SNP databases or an in-house reference database because they are unlikely to cause the disorder

Question 17

Exome comparison

Answer 17

Between different patients to find one or more affected genes in each of the patients (same variant is not required)

Question 18

Validation of cadidate genes

Answer 18

Wet-lab validation with Sanger sequencingfor example or comparison with sets of exomes and genomes

Question 19

Depending factors for stategy of indentifying causal alleles > impact sample size for adequate power in bioinformatics

Answer 19

-mode of inheritance (exome sequencing is more efficient for recessive disorders > less genes with two novel protein altering alleles)
-pedigree or population structure
-phenotype arising de novo or inherited > screening family
-extent of locus heterogeneity for a trait

Question 20

Filering data steps

Answer 20

1. discrete filtering: by comparing variants among individuals and against public databases/controls
2. Stratification of variants

Question 21

Novelty of allele assessment

Answer 21

-Set of public database polymorphisms like dbSNP and 1000 genomes project
> from unaffected individuals

Question 22

Filtering

Answer 22

Eliminating candidate genes by assuming any allele found in the filter set cannot be causative

Question 23

Assumption for filtering from dbSNP, and problems with the assumption

Answer 23

Controls do not have any alleles in the set from the individuals with the diseased phenotype
-Problems
>dbSNP is contaminated with a small number of pathogenic alleles
>some pathogenic alleles have a higher minor frequency: pathogenic gene variant also occurs in control exomes > risk of eliminating truly pathogenic alleles

Question 24

Stratification candidates (name the groups)

Answer 24

-by mutation type > predicted impact/deleteriousness
-by segmental duplications > variants found in segmental duplications are discarded
-by pseudognes: dysfuncitonal relatives of genes that have lost their protein-coding ability
-by function: predicted role of the protein product
-by functional impact: for non-synonymous alleles > impact on phenotype prediction

Question 25

(technical) Failure reasons of exome sequencing

Answer 25

-part of all the causative genes is not in the target definition
-inadequate coverage of the region which contains the causal variant
-the causal variant is covered but not accurately called
-false variants in a gene are called because of mismapped reads or alignment errors

Question 26

Failure with discrete filtering

Answer 26

reducing power due to genetic heterogeneity or false-positive calls (processed pseudogenes or segmental duplications)