Comparative genomics and Target discovery

1
Comparative genomics and Target discovery

• Maarten Sollewijn Gelpke
• MDI, Organon

2

• What is comparative genomics?
• What can we learn from comparative genomics?
• What is target discovery?
• What are the implications of comparative genomics
  to target discovery?
• What issues in target discovery can be addressed
  by comparative genomics?

3
Overview

• Introduction to genomes and sequencing.
• Comparative genomics aspects.
• Phylogenomics concepts.
• Examples of comparative genomics.

4
Sequence availability

• Availability of gene and protein sequences has
  increased enormously in during the last 2
  decades.
• Current capacity of the main sequencing centers
  is gt3Gb per month per centre.
• This will increase again dramatically with the
  development of new superfast sequencing
  techniques.

Currently gt 100Gbases
5
Genomes sequenced
A.thaliana
First bacterial genomes sequenced H.influenzae
and M.genitalium
The yeast genome
Human draft

• Human finished
• Rat
• Chicken

E.coli K12
Full sequence of chr. 22
1999
Xenopus Zebrafish
C.elegans
Chimpanzee
D.melanogaster Genome Chr. 21
6
Genome sequencing
Evolutionary relationship between metazoans
(multicellular animals) that have been sequenced
or are due for sequencing.
7
Genome sequencing

• BAC fingerprinting ? shotgun approach
• Accurate but laborious!
• Shotgun sequencing (WGS)

Bac
Clone
Bac
Clone
Whole Genome 30Mb 3Gb
100
-
200 kb
100
-
200 kb
Assembly
Finishing
8
Genome sequencing

• Current state of sequenced organisms
• gt316 Prokaryotes
• gt27 Archae
• gt280 Eukaryotes (complete, in assembly or in
  progress)
• gt1600 Viruses and gt 500 mitochondria/chloroplasts
• Some ongoing genome sequencing projects
• Poplar, gibbon, platypus, Drosophila species,
  variety of pathogenic fungi and bacteria, etc.
• Meta-genomic projects on environmental samples
  (soil, deep-sea, waste sites)

9
Future of genome sequencing?

• New complete genomes.
• New low-redundancy genomes.
• New (low-redundancy) genome areas.
• Meta-genomics. Sequencing of microbial
  communities.
• Sequencing of extinct species.
• 40000 year old Cave bear 26k, 21 genes.
• 45000 year old Neanderthaler 75k ? diverged from
  human lineage 315000 years ago

10
Comparative genomics

• Discover what lies hidden in genomic sequence by
  comparing sequence information.
• Main areas
• Whole genome alignment
• Gene prediction
• Regulatory element prediction
• Phylogenomics
• Pharmacogenetics
• Affected by evolutionary aspects
• Mutational forces (introduce random mutations)
• Selection pressures
• Ratio of non-synonymous to synonymous
  substitutions
• Mutation rates lower or higher than neutral

11
Comparing sequences, methods.

• Pairwise comparison of sequences (alignments)
• proteins or genes
• variety of local alignment tools like BLAST,
  Smith-Waterman etc.
• multiple sequence comparisons (ClustalW, Muscle
  etc.)
• results may be dependent on alignment settings

12
Comparing sequences, methods.

• Whole genome comparisons
• Large stretches of sequence
• Divergence up to 450Mya (fugu-human) with
  sufficient similarity remaining.
• BLAT, BLASTZ, Phusion/BlastN
• Seeding strategy ? alignment extension ? gapped
  alignments

13
Whole genome comparison

• Conservation of synteny!
• Cross-reference of any genetic traits (diseases!)
  from one organism (eg mouse) to genes in the
  syntenic regions in the other organism (eg
  human).
• Genome expansion and contraction
• Genome duplications, segmental duplications
  important mechanism for generating new genes.
• (GC) content, CpG islands
• Reflect different mutational or DNA repair
  processes?
• Repeats
• Transposable elements are a main force in
  reshaping genomes. TEs (or remainders thereof)
  can be used to measure evolutionary forces acting
  on the genome.
• Neutral mutation rate.

14
Gene prediction

• Comparing sequences has contributed enormously to
  the accuracy of gene prediction.
• Evidence based method.
• Use cDNAs, ESTs and proteins from various
  organisms.
• Apply gene feature rules.

15
Gene prediction

• De novo methods.
• Alignment of genomic sequences
• Splicing rules and other gene features
• De novo gene prediction by comparing sequences
  attempts to model a negative selection of
  mutations. Areas with less mutations are
  conserved because the mutations where detrimental
  for the organism.
• Prediction of similar proteins in both genomes.

Newly predicted protein in mouse and human,
similar to the disease related gene dystrophin.
16
Regulatory element prediction

• The complexity of higher eukaryotes and their
  relatively low number of genes can be explained
  partially through the importance of
  transcriptional regulation.
• Identification of REs will have an extensive
  impact in understanding gene expression patterns
  (expression intensity, tissue specificity),
  relations within expression patterns and
  inferring biological systems or networks.

17
Regulatory element prediction

• No formal models for regulatory motifs
• Attempt to find conserved regions or motifs based
  on the global alignment of similar sequences of
  different organisms (phylogenetic footprinting).
• Which species to compare? Evolutionary distance?
• What regions around gene models to investigate?
  5 and 3 flanking regions, introns?
• Take expression patterns into account?
• How does evolution affect REs?

18
Phylogenomics

• Comparison of genes and gene products across a
  number of species (whole genomes), characterizing
  homologues and gain insights in the evolutionary
  process itself.
• Pharmacophylogenomics is the use of phylogenomics
  in aid of drug discovery, through improved target
  selection and validation.

19
Orthology and paralogy
Phylogenetic tree of gene X

• Orthologs genes in different species that arose
  from a single gene in the most recent common
  ancestor, by speciation.
• Paralogs genes in the same species that arose
  from a single gene in a ancestral species, by a
  process of gene duplication.

20
Target orthology

• Species differences frequently affect progression
  of targets and compounds. Orthology maps in
  combination with expression studies may explain
  these differences.
• Establishing orthology
• Reciprocal highest scoring Blast hit.
• Conservation of synteny.
• Gene loss or rate of evolution issues.
• Orthology does not guarantee common function
  (functional shift).
• Extensive sequence divergence
• High non-synonymous over synonymous nucleotide
  substitution ratios.
• Comparison of regulatory regions?

21
Target paralogy

• Key insights in large pharmacologically relevant
  families (NRs, GPCRs) can be gained from paralogy
  analysis.
• Paralogy is inter-related with several other gene
  to function occurrences that can seriously affect
  the suitability of genes as drug targets

Schematic representation of various mappings of
genes to functions.
22

• Pleiotropy
• Suggested to precede paralogy
• Relaxed substrate or ligand specificity
• Multiple protein domains
• Tissue or cellular localization
• Redundancy
• Total or partial redundancy of function
• Directly linked to paralogy
• Robustness against gene knock-outs (target
  validation)
• PPAR-d / PPAR-a in skeletal muscle PXR / FXR in
  bile acid signaling dopamine transporters /
  serotonin transporters in adjacent neurons.

23

• Heteromery
• Formation of heteromers between paralogs
• Known examples in major classes of drug targets
• GPCRs GABAß receptors
• NRs formation of heterodimers with retinoid X
  rexeptors (RXR)
• Ion channels
• Crosstalk
• Combination of pleiotropy and redundancy
• May be regulated in time and space (expression
  and localization)
• Action of cytokines (interleukins) on immune cell
  types.

24

• Alternative transcription
• Intermediate between paralogy and pleiotropy.
  paralogy in place
• Increases effective size of the genome (estimated
  gt30 of human genes show alternative
  transcription!)

25
Effects on drug discovery

• Functional shifts, pleiotropy and redundancy
  potentially have good or bad news for drug
  discovery.
• Functional shifts
• Misleading or unavailable animal model
• Animal toxicity irrelevant for humans
• Pleiotropy
• Unintended drug effects
• Opportunities for multiple indications
• Redundancy
• Disease resistant to treatment (multi-functionalit
  y)
• Highly selective treatment for complex diseases.

26
Pharmacogenetics

• Within species comparative genomics
• Single Nucleotide Polymorphisms SNPs
• Current focus in coding regions, expected to
  expand to sites of transcription regulation.
• Determine the site of a SNP and the allele
  frequencies from ethnic or multi-ethnic panels of
  individuals (eg 100)
• Pharmacogenetics (PGx) relate SNP information to
  efficacy and safety issues during the drug
  development process.
• Efficacy PGx Select/predict drug responders,
  increase confidence in a certain drug in
  development.
• Safety PGx Identification of individuals with
  adverse effects to a drug

27
Examples

• New genes and REs from yeast genomes.
• Multi species comparisons from targeted genomic
  regions.
• Comparative genomics at the vertebrate extremes.
• Pharmacogenetics in drug efficacy

28
Comparison of yeast species to identify genes and
regulatory elements. (Kellis et al, Nature 2003)

• Saccharomyces cerevisiae and 3 related species
• 7x coverage WGS of each species
• Assembly of draft genome sequence
• S.cerevisiae genome aligned to others using ORFs
  as seeds
• Most ORFs have 11 matches. Considerable
  conserved synteny.
• Most genomic rearrangements clustered in
  telomeric regions.
• Local gene family expansion/contraction, creating
  phenotypic diversity over evolutionary time.
• Balance between conservation and divergence
  allows for accurate gene identification and
  recognition of REs as well!

29
Identification of genes

• Original S.cerevisiae genome (1996) 6275 ORFs
• Re-analysis and other evidence (2002) 6062 ORFs
• This study validates all ORFs using a reading
  frame conservation score (very sensitive).
• 5538 ORFs, 20 unresolved, 504 rejected ORFs!
• In addition to gene recognition, also largely
  improved gene structure definitions (start, stop,
  intron).

30
Identification of regulatory elements

• REs are difficult to identify
• Short (6-15bp), sequence variation, few known
  rules
• De novo discovery of REs directly from genomic
  sequence.
• Develop a motif conservation score system based
  on known motifs
• 78 motifs discovered, overlapping with 36 of 55
  known motifs
• Putative annotation of motifs using adjacent
  genes. (GO)
• 25 of 42 new motifs show high category annotation
  correlation
• Discovery of combinatorial control of Res
• Applications to human genome?
• Increase number of species in comparison to
  enrich the low signal to noise ratio in humans.

31
Multi species comparisons from targeted genomic
regions. (Thomas et al, Nature 2003)

• Comparing targeted regions areas in multiple
  evolutionary diverse vertebrates (less probable
  for conservation to occur by chance)
• ENCODE project
• 44 genomic regions (14 manually selected of
  which some disease related, 30 random) of diverse
  gene density and non-exonic conservation
• primates, bat, alligator, elephant, cat, emu,
  leopard, salmon etc.
• Initial analysis 1.8 Mb on chromosome 7
  containing 10 genes, including CFTR, from 12
  species.
• Detection of 1000 multi-species conserved
  sequences of which gt60 would not be detected by
  a 2 species comparison.

32
Comparative genomics at the vertebrate extremes
(Bofelli et al, Nature 2004)

• What can be learned from comparisons of genomes
  that are distant or closely related in evolution?
• Distant comparisons reveal the most constrained
  sequence elements.
• Most of the conserved human-fish non-coding
  sequences are found near genes with roles in
  embryonic development.
• Mutations can have an important role in human
  disease

33

• Human-Fugu conservation of non-coding sequence in
  the DACH gene area (development of brain, limbs,
  sensory organs).
• Validation of identified enhancer regions by
  driving expression of a reporter in mouse embryos.

34
Comparative genomics at the vertebrate extremes

• Intraspecies sequence comparisons allow
  identification of species specific sequences
• Phylogenetic shadowing
• Requires high rate of polymorphism
• Comparison among primates show human specific
  sequences
• Analysis of regulatory sequence of ApoA (involved
  in human heart disease)

A. Mutation rate analysis of Ciona intestinalis
5 region of the forkhead gene. B. Validation
of identified potential regulatory elements in
Ciona larvae.
35
Pharmacogenetics in drug efficacy
Efficacy PGx for an obesity drug. Compare
genotypes 1-1, 1-2 and 2-2