Inside the Genome

1
Inside the Genome
2
2001 The Human Genome
The club resident JD Watson Back2back with DJ.
Venter and
International Human Genome Sequencing
Consortium, Nature, 409 860-921 (2001)
Venter et. al. , Science 2921304-1351 (2001)
3
Prologue RNA word the dark matter of genomics

• How many coding genes in the human genome?
• The Bet of 2000
• Mean 61710
• Range 30,000 150,000
• By the end of the genome project the estimated
  number of human protein-coding genes declined to
  only 25,000
• What is the source for that discrepancy?
• ESTs based estimation Vs. Whole Genome annotation

4
RNA revolution

• The majority of the transcriptional output comes
  from non coding RNA
• an average of 10 of the human genome (compared
  with 1.5 exonic sequences) resulted in
  transcripts Cheng et al. 2005
• Or even more...62 of the mouse genome is
  transcribed FANTOM3 Science 2005

5
Various RNAs A partial list

• messenger RNA (mRNA)
• Ribosomal RNA (rRNA)
• Transfer RNA (tRNA)
• Small nuclear RNA (snRNA)
• Small nucleolar RNA (snoRNA)
• Short interfering RNA (siRNA)
• Micro RNA (miRNA)

6
RNAs are not merely the intermediary cousins of
proteins -The Central dogma of molecular biology
Revisited
Genome
miRNA
Regulation by proteins
Regulation by RNA
Transcriptome
Proteome
7
Research in Biology is complex

• Deciphering Biological Systems
• The advantage (what makes this quest feasible)
  and the hindrance (what makes this quest
  inherently difficult) both explained by
  evolution.

8
The Hindrance Topological Entanglement of
functional interconnections

• The difficulties in our research fundamentally
  owe their complexity to the designer natural
  selection.
• What is it - a Robot or a UFO ?
• The reason lies in the profound difference
  between systems designed by natural selection
  and those designed by intelligent engineers
  Langton 1989 Artificial Life.

9

• Bottom linewe investigate an outrageously
  complex weave of interconnections
• The textbook networks represent only the tip of
  the iceberg.
• miRNAs and Regolomics
• microRNAs - Expected to represent 1 of
  predicted genes Lim et al., 2003
• Lewis et al., (2003) estimate average of five
  targets per miRNA
• Many targets are transcription factors - miRNAs
  regulate the regulators

10
The advantage universal homology, thus
enabling comparative biology.

• Bottom linethe research in biology advances
  through a reductionist approach - using simple
  model organisms to infer functionality of
  homologous systems.

11
Human genome statistics
2.91 billion base pairs 24,000 protein coding
genes (gt30,000 non-coding genes ???) 1.5 exons
(127 nucleotides) 24 introns (3,000
nucleotides) 75 intergenic (no genes) Repetitive
elements rule ( 45 dispersed repeat) Average
size of a gene is 27,894 bases Contains an
average of 8.8 exonsTitin contains 234
exons. Ave. of 4 diff. proteins per gene
(alternative splicing)
12
Detecting genes in the human genome

• Gene finding methods
• Ab initio use general knowledge of gene
  structure rules and statisticsThe challenge
  small exons in a sea of introns
• Homology-based The problem will not detect
  novel genes

13
Genscan (ab initio)
\\// (o o) -. .-.
.-oOOo(_)oOOo-. .-. .-. .-. .-. .-. .-.
.-. .-. .-. .-. .-. .-. .-. .-. .-. .-.
X\ /X\ /X\ /X\ /X\
/X\ /X\ /X\ /X\ /X\
/X\ / \X/ \X/ \X/
\X/ \X/ \X/ \X/ \X/
\X/ \X/ \ ' -' -' -' -'
-' -' -' -' -' -' -' -' -' -'
-' -' -' -' -' -' -'

• Based on a probabilistic model of a gene
  structure
• Takes into account- promoters - gene
  composition exons/introns- GC content- splice
  signals
• Goes over all 6 reading frames

Burge and Karlin, 1997, Prediction of complete
gene structure in human genomic DNA, J. Mol.
Biol. 268
14
Splicing
15
Eukaryotic splice sites
Poly-pyrimidine tract
16
CpG Islands another signal

• CpG islands are regions of the genome with a
  higher frequency of CG dinucleotides (not
  base-pairs!) than the rest of the genome
• CpG islands often occur near the beginning of
  genes ? maybe related to the binding of the TF
  Sp1

17
Gene Ontology

• GO describes proteins in terms of biological
  process(e.g. induction of apoptosis by external
  signals) cellular component(e.g. membrane
  fraction)molecular function(e.g. protein
  kinase)

18
Comparative proteome analysis
Functional categories based on GO
19
Comparative proteome analysis

• Humans have more proteins involved in
  cytoskeleton, immune defense, and transcription

20
Evolutionary conservation of human proteins
???
21
Horizontal (lateral) gene transfer

• Lateral Gene Transfer (LGT) is any process in
  which an organism transfers genetic material to
  another organism that is not its offspring

22

• Mechanisms
• Transformation
• Transduction (phages/viruses)
• Conjugation

23
Bacteria to vertebrate LGT detection

• E-value of bacterial homolog X9 better than
  eukaryal homolog

Human query Hit e-value Frog ..
4e-180 Mouse 1e-164 E.Coli .. 7e-124
Streptococcus .. 9e-71 Worm .0.1
24
Bacteria to vertebrate LGT
Non-vertebrates
Bacteria
vertebrates
25
(No Transcript)
26
Bacteria to vertebrate LGT??

• Hundreds of sequenced bacterial genome vs.
  handful of eukaryotes
• Gene finding in bacteria is much easier than in
  eukaryotes
• On the practical side rigid mechanical barriers
  to LGT in eukaryotes (nucleus, germ line)

27
Repetitive Elements in the Human Genome
28
Repeats statistics

• The human genome is 45 dispersed repeat
• 20 LINEs, (AT rich)
• 13 is SINES (11 Alu), (GC rich)
• 8 LTR (retrovirus like) and
• 2 DNA transposons
• Another 3 is tandem simple sequence repeats
  (e.g. triplet)
• And another 3-5 is segmentally duplicated at
  high similarity (over 1kb over 90 id)
• Identifying and screening these out is essential
  to avoid fake matches

29
LINEs and SINEs

• Highly successful elements in eukaryotes
• LINE - Long Interspersed Nuclear Element (gt5,000
  bp)
• SINE - Short Interspersed Nuclear Element (lt 500
  bp)
• SINEs are freeriders on the backs of LINEs
  encode no proteins

30
The C-value paradox

• Genome size does not correlate with organism
  complexity

Amoeba Rice Human Yeast
670 billion 4.3 billion 3 billion 12 million Genome size
? 30,000 20-25,000 6,275 Number of genes
31
Repetitive elements

• The C-value mystery was partially resolved when
  it was found that large portions of genomes
  contain repetitive elements

32
Are Alus functional??

• SINEs are transcribed under stress
• SINE RNAs may bind a protein kinase ? promote
  translation under stress
• Need to be in regions which are highly
  transcribed
• Role in alternative splicing

33
Segment duplications

• 1077 segmental duplications detected
• Several genes in the duplicated regions
  associated with diseases (may be related to
  homologous recombination)
• Most are recent duplications (conservation of
  entire segment, versus conservation of coding
  sequences only)

34
Genome-wide studies
35
Sequenced genomes
36

• 481 segments gt 200 bp absolutely conserved (100
  identity) between human, rat and mouse

37
Comparison with a neutral substitution rate

• Compare the substitution rate in a any 1Mb region
• Probability of 10-22 of obtaining 1
  ultranconserved element (UE) by chance

38
481 UEs
100 intronic
111 UE overlap a known mRNA exonic UEs
256 - no overlap (non-exonic)
156 inter-genic
114 - inconclusive
39
Who are the genes?
Type 1 exonic Type 2 genes which are near
non-exonic UEs (???)
40
Intergenic UEs

• Genes which flank intergenic UEs are enriched for
  early developmental genes
• Are UEs distal enhancers of these genes?

41
Gene enhancer

• A short region of DNA, usually quite distant from
  a gene (due to chromatin complex folding), which
  binds an activator
• An activator recruits transcription factors to
  the gene

42
Experimental studies of UEs
Tested 167 UEs (both mouse-human UEs and
fish-human UEs) for enhancer activity cloned
before a reporter gene to test their
activity 45 functioned as enhancers
43
A bioinformatic success

• Ultraconservation can predict highly important
  function!

44
BUT
Ahituv PLoS Biol. 2007 Sep5(9)e234
Chose 4 UEs which are near specific genesgenes
which show a specific phenotype when
knocked-out Performed complete deletion of these
UEs the mice were viable and did not show any
different phenotype
45
Conclusions

• Ultraconservation can be indicative of important
  function
• And sometimes not- gene redundancy- long-range
  phenotypes- laboratories cannot mimic life