Lecture 9 - Coding Vs Non Coding Variation

Question 1

What is the most common human genetic variation?

Answer 1

Human polymorphism

SNP is Single nucleotide polymorphisms —> single nucleotide substitutions is most common

• Human genome has millions of SNPs

Question 2

Most types of SNP? Their affect?

Challenge?

Answer 2

Most SNPs are NEUTRAL = no phenotype

Most SNPs are in NON-CODING REGIONS

—> Non-coding SNPs are likely more IMPORTANT IN COMPLEX DISEASES AND PHENOTYPES

The CHALLENGE: Which of the variants alter phenotype?

Question 3

Types of Functional Variants : 2

Answer 3

1. CODING variation
2. Non coding/ regulatory variation

Question 4

Types of Functional Variation: CODING -2

Answer 4

• Amino acid variation

• Splice/reading frame variation

Question 5

Types of functional variation: Non - Coding Variation - 4

Answer 5

• Transcriptional

• Post-transcriptional (mRNA processing)

• Non-coding RNA (miRNA, lncRNA)

• Epigenetic

Question 6

Types of functional variation: Non - Coding Variation - 4

Answer 6

• Transcriptional

• Post-transcriptional (mRNA processing)

• Non-coding RNA (miRNA, lncRNA)

• Epigenetic

Question 7

Every locus in Unqiue …

Answer 7

Every locus unique - requires a different approach

Question 8

Overview of Functional annotation

Answer 8

A.
- GWAS
- GENOME
- EXOME
- Linkage and sequence analysis

B. Sequencing on Chromosome

C. Variants of various functional classes

D. Comparative genomics

E. Biochemistry/ Structure

F. Experimental function

Question 9

Overview of Functional Enrichment

Where are the Regulatory elements?

Answer 9

Most functional variation is NON CODING AND REGULATORY

Genome
- coding 0.5%
- UTR 0.8%
- promoter 2.2%
- DHS 15.7%
- enhancer 3.2%
- intergenic DNA 52.0%
- Introns 28.8%

Question 10

Workflow for Variant Functional Identification - 5

Answer 10

1. FINE MAPPING
2. In silico annotation
3. SNP function
   4 . Target gene(s) identification
4. Target gene function

Question 11

Fine mapping involves:

Answer 11

• Locus genotyping (sequencing)

• Statistical tests for independent associations

• LD mapping

Question 12

In silico annotation:

Answer 12

• Integrative tools (ENCODE)

• Exome-translation

• Intron-intergenic

Question 13

SNP function —> target gene/s identification

Answer 13

• Coding– GERP , PolyPhen
  • Non-coding– GERP, RegulomeDB (ENCODE)➢ Transcr. factor element –EMSA, reporter gene, CHA-seq
  ➢ miRNA - TargetScan, miRanda
  ➢ lncRNAs– REMSA
  ➢ Epigenetic variants - MeQTL

Question 14

Target gene/s function

Answer 14

• Cell culture models, human tissues

• Isogeneic models (CRISPR/Cas)

• Animal models– mouse KOs, zebrafish, Drosophilia

Question 15

Identification of Coding-Region Variants: 3 types

Answer 15

1. Frame-shift variation (indels) etc —> protein truncations, fusions
2. Base-substitution: 2 types
3. Splice-site variation
   - (may change AA sequence, may have
   non-coding effects e.g. regulation of RNA processing)

Question 16

2 types of Base-substitution: 5 points

Answer 16

1. synonymous– base change with no amino acid change
2. non-synonymous– base change with AA change
   … 3. Conservative– change to similar AA

… 4. Semi-conservative – e.g.
-ve to +ve charged AA

… 5. Radical – AA with very different properties

Question 17

Approaches to predicting Coding Variant Functions:

Tools for nucleotide-sequence-based prediction of deleteriousness

Answer 17

GERP : single site scoring - evolutionary

Question 18

Approaches to predicting Coding Variant Functions

Tools for protein-sequence-based prediction of deleteriousness

Answer 18

Polyphen: trained classifier : evolutionary, biochemistry and structural

Question 19

Bioinformatics approaches to predicting coding variant function:

Polyphen vs GERP

Answer 19

‘Genomic Evolutionary Rate Profiling (GERP)’
- ORTHOLOGOUS nucleotide sequences COMPARED to DETERMINE EVOLUTIONARY CONSTRAINTS TO CHANGE IN SEQUENCE

‘Polyphen’
- PREDICTS impact of AA SUBSTITUTION on STRUCTURE & FUNCTION of a
PROTEIN
- using PHYSICAL and COMPARATIVE CONSIDERATIONS

Question 20

Genomic Evolutionary Rate Profiling (GERP) in detail 5.

Answer 20

1. leverage comparative NUCLEOTIDE sequence information by looking for REGIONS THAT EXHIBIT EVIDENCE OF SELECTIVE CONSTRAINT
2. IDENTIFIES CONSTRAINED ELEMENTS (strings of nucleotides) by QUANTIFYING SUBSTITUTION DEFICITS
   ‘i.e. deficits represent substitutions that would have occurred if the
   element was neutral DNA, but didn’t occur due to selective pressure’
3. Remember: Conservation equals function
4. • GERP ESTIMATES CONSTRAINTS FOR EACH ALIGNMENT COLUMN COMPARED TO NO CONSTRAINT
5. R = sum (expected – observed rate)

Question 21

PolyPhen: Polymorphism Phenotyping

Answer 21

1. Predicts impact of AA SUBSTITUTION on the STABILITY and FUNCTION OF A PROTEIN.
2. Uses PHYSICAL (3-D structure) & COMPARATIVE EVOLUTIONARY COMPARISONS
3. Estimates PROBABILITY of VARIANT BEING DAMAGING TO PROTEIN FUNCTION OR STRUCTURE.
4. ‘Prediction outcome’ - PROBABLE DAMAGING, POSSIBLY damaging, or BENIGN
5. PolyPhen-2 found at a website.

Question 22

What is ENCODE? what does it do?

Answer 22

1. a project to identify ALL FUNCTIONAL ELEMENTS IN THE HUMAN GENOME
   SEQUENCE.
2. Transcription factor binding sites (ChIP-seq)
3. DNase I Hypersensitive sites (DNase-seq)
4. regRNAbinding sites
5. SNP catalogue (1000 genomes, GWAS)

Question 23

Prioritisation scores - RegulomeDB

Answer 23

lower scores indicate increasing evidence for a variant to be locate in a functional region.

category 1 variants have equivalents in other categories with the additional requirment of eQTL information.

Question 23

Bioinformatic Identification of Regulatory Variants:
‘Regulome DB’

Answer 23

1. INTEGRATES FUNCTIONAL DATA contained in ENCODE with GENETIC VARIATION DATABASES
   (dbSNP, ClinVar etc)
2. Predicts WHETHE VARIANTS are FUNCTIONAL - PRIORITISATION SCORE.

Question 24

example SNV rs9261424 overlapping many regulatory.

Answer 24

NFKB track for 3 individuals:
1. homozygous to
reference allele (G),

2. heterozygous, and
3. homozygous to alternate allele (C)

Question 24

Summary – Lead SNP to Functional SNP

Answer 24

1. Genome
2. Predicted motifs
3. DNAseI Hypersensitivity Peaks
4. ChIP-seq Peaks for TF1
5. Linkage Disequilibrium

Question 25

What’s Next?

Answer 25

What’s Next?
“A major goal will be to develop a unified, quantitative, predictive
framework to estimate the prior probabilities for any given mutation to be
both functionally relevant and disease-relevant, accounting for both
computational and experimental sources of information.

A number of
challenges must be met for such a framework to succeed”

• need large collections of true positive (functional) and true
  negative (neutral) variant