Genome Sequence Acquisition and Analysis Lecture3

1
Genome Sequence Acquisition and AnalysisLecture-3
2
Data Acquisition

• Sequencing the Whole Genome
• Sequencing technology
• Polymerase chain reaction leads to cycle
  sequencing.
• Capillary electrophoresis
• Sequencing strategies for mapping whole genomes
• Hierarchical Shotgun Sequencing
• Shotgun Sequencing
• Assembling strategies of the first draft of the
  Human genome
• How do we make sense of all these bases?

3
Reading

• Chapter II,III, XIII Baxevanis

4
Automated DNA sequencing

• DNA polymerase copy a strand of DNA
• The insertion of dye terminator stop the DNA
  synthesis. The different fragments are run on a
  gel one lane contain each 4 bases terminator
• The terminator are labeled with different dye
  depending whether it is A,G,C,or T
• The sequence is read by a computer. It generate
  a sequence trace, the computer translate the
  fluorescence signal into DNA sequence.

5
Thermocycler for PCR and sequencing
6
Automated cycling DNA sequencing
Terminator Dye

• Using a DNA polymerase thermostable, it is
  possible to use the polymerase chain reaction
  method for sequencing. The advantage is that we
  can generate sequencing from very little amount
  of DNA since the template is amplified during the
  process.

Polymerase
550C anealling of the
Primer with DNA
720C
DNA synthesis Of different fragments With
different dye Terminator and length
950C
Dissociation of Duplex DNA amplification Of the
template
7
Capillary Electrophoresis
Electrophoresis is a powerful method that was
developed long before molecular biology came into
existence. The principle behind electrophoresis
is that charged molecules will migrate toward the
opposite pole and separate from each other based
on physical characteristics. For example,
SDS-PAGE separates proteins according to their
molecular weights. The two limiting factors of
traditional electrophoresis were 1) detection of
molecules upon completion of electrophoretic
separation and 2) only low voltages could be used
to prevent heat damage of the samples. Capillary
electrophoresis solved both of these problems.
Because the capillary tube has a high surface to
volume ratio (25-100 µm diameter), it radiates
heat readily and thus samples do not over heat.
Detection of the migrating molecules is
accomplished by shining a light source through a
portion of the tubing and detecting the light
emitted from the other side (figure 1).
Scanner
8
The Mega-sequencer
9
The Mega-sequencer
10
Data acquisition
11
Mapping strategies
12
Definitions

• BAC Bacterial Artificial Chromosomes are used to
  clone large fragments of DNA up to 150kb and can
  replicate in E.Coli. Insert are shorter than the
  YAC vector. They are the most used today for
  hierarchical sequencing.
• YAC Yeast Artificial Chromosomes are used to
  clone large fragments of DNA for sequencing from
  150kb to 1.5Mb. The problem is that the insert
  are sometime instable and rearrangment of the DNA
  can occur.
• STS To allow better assembling of the different
  fragment obtained from sequencing of a genome,
  sequence tag sites were mapped all other the
  human genome. They are short unique sequences of
  DNA all along the chromosomes and defined by a
  pair of primers that amplify only one segment of
  the chromosome.
• EST When the genomes started to be annotated
  means define the coding region of the genes,
  several labs started to sequence short cDNA
  transcript lt500bp from either 3 or 5 end of the
  DNA. The short segments of DNA were called
  Expressed Sequence Tags (EST), it did not matter
  if they knew from which genes they came since
  their goal was to compile every EST. EST have
  been very useful in cloning genes identify
  splicing variants of genes. If investigator
  sequenced a piece of DNA that contained an EST,
  they would know that the DNA would be a coding
  region. Those EST grown very rapidely and are
  now organized in an EST database called unigene.
• Today many sequencing consortium are dayly
  producing full length cDNA that are added to
  UNIGENE database the goal is to have all the full
  length cDNA of all the genes available for the
  research community. These genes can be bought
  for (300US)
• FISH Fluorescence in situ hybridization.
  Chromosomes are isolated, fixed on a glass slide
  and hybridized to a fluorescent molecular probe
  this experiment allow the position of markers
  (STS, BAC) onto the cytogenetic map of the
  chromosome.
• CONTIGS BAC and YAC vectors contains contiguous
  fragments of DNA that overlap

13
BACs

• Large insert size
• No Chimeras
• Stable in host cells

BAC ends
14
Chromosome Map Cytogenetic Map
Centromere
Distant Markers
Closely linked markers
BAC Fragment collection Physical Map
Mostly found on same fragments
Rarely found on same fragments
Phred and Phrap (Phil Green) suite of programs
are used to assemble the DNA fragments and
assesss the quality of the assembly. (Baxevanis
Chapter XIII).
15
Feature annotationFluorescence in Situ
HybridizationFISH

• When a particular marker (STS, BAC) is positioned
  on the cytogenetic map of a chromosome by FISH
  all DNA fragments (BAC) that contain the marker
  and are overlapping will be ultimately positioned
  on the chromosome to build the physical map

An example of Fluorescence in situ hybridization
using a BACs subsequences as a fluorescent
probe
16
Physical Map assembly

• Library of BACs containing genome fragments
• STS linkage to position fragments
• Physical Map assembly
• Fragment BACs and assemble sequenced contigs
• Align sequenced fragments to physical map by
  Phred and Phrap.

17
Hierarchical shotgun sequencing
The method preferred by the Human Genome Project
is the hierarchical shotgun sequencing method. In
this approach, genomic DNA is cut into pieces of
about 150 kb and inserted into BAC vectors,
transformed into E. coli where they are
replicated and stored. The BAC inserts are
isolated and mapped to determine the order of
each cloned 150 kb fragment. This is referred to
as the Golden Tiling Path. Each BAC fragment in
the Golden Path is fragmented randomly into
smaller pieces and each piece is cloned into a
plasmid and sequenced on both strands. These
sequences are aligned so that identical sequences
are overlapping. These contiguous pieces are then
assembled into finished sequence once each strand
has been sequenced about 4 times to produce 8X
coverage of high quality data.
18
Shotgun sequencingCelera method

• The method developed and preferred by Celera is
  simply called shotgun sequencing. This approach
  was developed and perfected on prokaryotic
  genomes which are smaller in size and contain
  less repetitive DNA. Shotgun sequencing randomly
  shears genomic DNA into small pieces which are
  cloned into plasmids and sequenced on both
  strands, thus eliminating the BAC step from the
  HGP's approach. Once the sequences are obtained,
  they are aligned and assembled into finished
  sequence.

19
Which strategy is the best

• When the human project were published in nature
  2001(HGP academic) and science (celera) many of
  the contigs were not overlaping creating gaps in
  the genome.
• Many believe that the HPG method is better suited
  for filling gap between the contig since the
  fragments are ordered.
• For example duplicated genes could be combined in
  a single loci by the celera method, while the HPG
  method could identify two different loci from the
  duplicated genes.
• The completed human genome by the HPG has been
  released in april 2003 for the 50th anniversary
  of the double helix. Celera has been able to
  update their genome using the HPG genome. The
  human genome produced by the HPG consortium did
  not took advantage of the celera genome since
  celera did not share their data for intellectual
  property reasons.
• Today the HPG genome is available on public
  database, while the celera genome is only
  available by subscription (10,000 USD)

20
Assembling the Human Genome

• The HPG was not limited to the human genome it
  also targeted,fly,worm,mouse,plants,and yeast.
  By comparing different genomes we should be able
  to learn more about genome in general and human
  genome in particular
• Big genomes contain repetitives regions from old
  viruses and transposons that have been
  accumulated during millena. Piecing together the
  full genome would be better if some markers
  existed. For this purpose STS were designed to
  position the different pieces on the genome.
• Since many regions of the DNA do not code for RNA
  or proteins, EST were used to differentiate
  between junk DNA and DNA coding regions.

21
Assembling the Human Genome

• Assemble the sequence Data
• Annotate the features
• Provide a dataset of assemble genomic
  sequence,RNAs, and proteins
• Create a mapviewer of the genomic sequences

22
Genomics sequence assemblies

• Public effortObjectives Assemble overlapping
  fragments from different BAC sequencesTwo
  organizations
• The Human Genome Project Working Draft (also
  known as the 'Golden Path' assemblies) at
  University of California, Santa Cruz (UCSC)
• The National Center for Biotechnology
  Information (NCBI) Contig Assemblies.

23
Genomics sequence assemblies

• Public effortThe UCSC strategy begins with human
  genomic sequences from GenBank composed of BAC
  clones at a given point (a 'freeze' dataset),
  ordered and oriented according to the appropriate
  fingerprinting contigs from Washington University
  Genome Sequencing Center (WUGSC) were obtained.

-
A typical fingerprint gel containing 121 lanes of
DNA. Every 5th lane is a marker lane with 37
bands to calibrate the gel metric. Each of the 96
data lanes contains a digested BAC, with 30-60
bands.

24
Genomics sequence assemblies

• Public effort Within each WUGSC contig, draft
  sequence fragments are assembled into consensus
  'raft' sequences using overlaps between fragments
  and bridging mRNA, expressed sequence tag (EST),
  plasmid and BAC-end-pair sequences. Repeated
  tracts of the letter 'N' are inserted between
  non-overlapping rafts to give a longer consensus
  sequence for each WUGSC contig.

25
Genomics sequence assemblies

• Public effortThe NCBI approach also begins by
  finding an order for adjacent BACs but in this
  case it is derived from BAC sequence overlaps
  chromosome assignment by fluorescence in situ
  hybridization (FISH) and sequence-tagged site
  (STS) content. The sequence fragments from these
  overlapping BACs are then merged into consensus
  'meld' sequences. As with the UCSC method, these
  melds are then ordered and orientated using ESTs,
  mRNAs, etc.., before being combined into a single
  NCBI genomic sequence contig with melds separated
  by runs of the letter 'N'. Because NCBI will
  often be dealing with fewer sequence fragments
  than UCSC in the construction of a given contig
  (since NCBI only assemble contigs from
  overlapping BAC sequences, not from all BAC
  sequences within an FPC contig as does UCSC,
  there should be less opportunity for misassembly
  (erroneously ordering or orientating sequence
  fragments).

26
Genomics sequence assemblies

• Private effortCelera Genomics has also published
  a draft assembly for the genome. It is available,
  under a variety of restrictions, only from their
  web site.
• They used a whole-genome shotgun sequencing
  approach to obtain fragmentary and unmapped
  sequence data, and combined these with
  International Human Genome Sequencing Consortium
  (IHGSC) sequence and mapping data.
• Celera assembled its own data with those produced
  by the IHGSC using two different assembly
  protocols either with or without mapping
  information. They found that using the public
  mapping data gave the assembly greater sequence
  coverage.
• The Celera sequences may provide a useful
  resource for plugging gaps in the IHGSC draft
  genome. There were up to 0.14 of Celera
  sequence was not present in the public data set
  in the first draft of the human genome.

27
Genomics sequence assemblies

• The UCSC effort provided an exhaustive account of
  the shortcomings of their assemblies. Summary
  statistics are given for almost every measurable
  characteristic, including numbers and lengths of
  gaps and the estimated number of misassemblies.
• The UCSC group has examined the performance of
  their assembly algorithms on artificially
  produced fragmentary sequence (fragmented
  sequences from finished regions of the genome).
• Their assemblies reproduced the correct ordering
  of fragments around 85 of the time and the
  correct orientation of fragments around 90 of
  the time.
• The quality of the assemblies degraded rapidly
  where the algorithms had to deal with many small
  fragments, with both correct order and
  orientation being achieved only 50 of the time
  (effectively the same as random assembly) in the
  worst cases.
• It was also identified possible misassemblies in
  the Celera assembly. Neither Celera nor NCBI
  publishes statistics analogous to UCSC's for
  draft sequence assembly, nor have they published
  estimates of the rate of error involved.

28
Genomics Mapping Data

• WUGSC(Washington University Genome Science
  Center) has produced a physical map of the
  genome using 'fingerprinting' analysis of BACs
  from a library of genomic sequence clones called
  RPCI-11
• Overlaps between clones are calculated using
  different algorithms on the basis of clone
  restriction-fragment patterns, or fingerprints.
  In this way, fingerprinting-based contigs
  covering at least 96 of the euchromatic genome
  have been constructed. The first published draft
  genome sequence assemblies contain hundreds of
  thousands of gaps, whereas this physical map
  contains less than 1000.
• The International Human Genome Mapping Consortium
  (IHGMC) physical map were supplemented with data
  from other sources. Many fully or partially
  sequenced BAC clones have been cytogenetically
  mapped to particular regions using FISH, and this
  cytogenetic location is given in the GenBank
  sequence entry for the clone.
• The end sequences from many BAC clones, including
  those from libraries other than RPCI-11, are
  available from The Institute for Genomic Research
  (TIGR) Human BAC Ends site.
• The physical and genetic marker content of a
  sequence of interest can be determined online
  using the electronic-PCR (e-PCR) program at NCBI.
  This is a rapid sequence-search algorithm that
  searches your sequence for occurrences of marker
  sequences in GenBank. If you enter a BAC clone
  accession number, the output consists of an
  ordered list of markers (and the chromosome they
  map to) down the clone sequence.

29
Possible stage of BAC clone sequence

• Phase 1 Unfinishedunordered/unoriented contigs
  with gaps
• Phase 2 Unfinished ordered/oriented contigs
  with or without gaps
• Phase 3 Finished no gaps

30
Writing the parts listgenomic sequence
annotation

• "We've called the human genome the blueprint, the
  Holy Grail, all sorts of things. It's a parts
  list," said Eric Lander at the Millennium Evening
  at the White House, 14 October 1999 about the
  first draft of the human genome.
• The IHGSC process of predicting genes were
  estimated to have a sensitivity of 68-85 (so
  that 15-32 of genes present in the genome were
  missed) and an accuracy of around 79 on average
  (so that 21 of the sequence of each predicted
  gene was missed).
• Since then gene prediction has been constantly
  improved

31
Writing the parts listgenomic sequence
annotation

• To obtain a better gene prediction in the human
  genome cDNA libraries were made and screened to
  verify or refute gene predictions made for the
  first completed human chromosomes, at the Sanger
  Centre.
• Many genes from a large collection of full-length
  mouse cDNAs and large human cDNA collections were
  produced.
• The IHGSC has assembled an initial, non-redundant
  set of predicted and known human genes called the
  integrated gene index (IGI), which is to be
  updated as more data accumulate and is made
  accessible via the Ensembl site.
• Beyond the question of where exactly the genes
  are, it would also be useful to have some idea of
  what they do.
• Only around 40 of the IGI-predicted proteins
  appeared initially to contain known domains
  implying that we may be able to predict at least
  some of their functions and even in these cases
  the computational predictions must be supported
  by lab work before they are regarded as accurate.

32
Genomes Browsers

• There are three major well-designed websites
  offering users the chance to browse annotations
  of the human genome.They offer a graphical
  interface to display the results of various
  analyses, such as gene predictions and similarity
  searches, for draft and finished genomic
  sequence.
• They allow rapid intuitive comparisons between
  the features predicted by different programs.
• One can see at once where an exon prediction
  overlaps with interspersed repeats or a
  single-nucleotide polymorphism (SNP).
• There are differences between the 3 different
  Browser.

33
Ensembl

• The Ensembl database was the first to provide a
  window on the draft genome and started by
  curating 'confirmed' genes that were
  computationally predicted and also supported by a
  significant match to one or more expressed
  sequences.
• Uses gene structures from public sequence
  database entries (many of which are
  experimentally verified), so the total set of
  Ensembl genes should be a more accurate
  reflection of reality than computational
  predictions alone.
• Other features includes repetitive sequence,
  cytological bands, genetic markers, CpG island
  predictions, tRNA gene predictions, UniGene
  (expressed sequence) clusters, SNPs from the
  dbSNP database, disease genes found in the draft
  genome (identified by the Online Mendelian
  Inheritance in Man database, OMIM, and regions of
  homology to mouse draft genomic sequences.
• Ensembl uses the UCSC draft sequence assemblies
  as its starting point, so its description of a
  region can only be as accurate as those
  assemblies allow. Gaps and misassemblies in the
  genomic sequence could lead to Ensembl missing or
  wrongly positioning genes and other features.
• Misassemblies are expected to arise most
  frequently among small BAC fragments, but more
  broadly the BAC clones mapped to a given region
  by UCSC (on the basis of FPC data, FISH data, STS
  content and sequence overlap) should be very
  accurate. Thus, Ensembl annotation may not be
  accurate on the finest, local scales but should
  give a very good idea of what is present in a
  larger region, perhaps at the level of megabases.

34
UCSC Human Genome Browser

• The UCSC Human Genome Browser (HGB Figure 2) 4
  bears many similarities to Ensembl
• it too provides annotation of the UCSC assemblies
• it displays a similar array of features.
• Additional features of HGB not yet found in
  Ensembl includes.
• Predictions from more than one ab initio
  gene-prediction program (programs that predict
  coding sequence on the basis of statistical
  measures of features such as codon usage,
  initiation or polyadenylation signals, rather
  than by homology to known genes) and indicates
  regions with significant homology to other
  organisms.
• These features can provide useful information
  when dealing with gene predictions that are not
  well supported by similarity to known mRNA
  sequences.
• Detailed description of the genomic sequence
  assemblies. Graphical representations of the
  fragments making up a region of draft genome can
  be displayed, showing the relative size and
  sequence quality of each fragment and also
  whether any gaps between fragments are bridged by
  mRNAs or paired BAC end sequences.ne can get an
  idea of the likely degree of misassembly in a
  region. Data retrieval is possible via text, BLAT
  searches (a faster, less accurate algorithm than
  BLAST) and bulk downloads of annotation or
  sequence data.

35
NCBI Map Viewer

• Whereas Ensembl and HGB both show annotation of
  the UCSC draft genome assemblies the NCBI Map
  Viewer (NMV) displays features present in the
  NCBI assemblies. The NMV shows
• Comparisons between cytogenetic, genetic and
  radiation hybrid maps in parallel with NCBI draft
  and finished sequence contigs. The locations of
  genes, STSs, and SNPs are indicated on the contig
  sequences. The NCBI approach to gene prediction
  is more conservative than those in Ensembl and
  HGB.
• No ab initio gene-prediction-program is used
  instead of known genes, mRNAs and ESTs are
  aligned to genomic sequence using a program
  called Acembly. The program also attempts to give
  alternative splice variants of genes where its
  alignments suggest them. All annotated genes are
  connected to NCBI LocusLink, which provides links
  to associated information such as related
  sequence accession numbers, expression data,
  known phenotypes, polymorphisms, and so on.

36
NCBI Map Viewer

• In spite of difficulties with the quality of
  genomic sequence assemblies, the three browsers
  remain extremely useful tools for the cautious
  biologist. They undoubtedly indicate the presence
  of most of the coding sequence in a given
  fragment of genomic sequence and indicate its
  location in the genome as determined by the best
  available mapping data.
• Most aspects of the analysis carried out are the
  subjects of active research, and improvements in
  performance resulting from the inclusion of new
  sequence data and algorithms will be ongoing
  annotations are constantly subject to changes.

37
End Product

• The NCBI annotation project provide sequences and
  resources support via AceView,LocusLink and
  Mapviewer.
• RefSeq provides accession prefixes for each of
  the type of sequences
• XM_XXXmRNA
• XR_XXXXnon-coding transcript
• XP_XXX Protein, model reference sequence
• The NCBI equivalent accession number on the NCBI
  contigs server are NT_ or NW which may be
  generated from incomplete data with mistakes.

38
NCBI Assembly Pipeline
39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
Useful Websites

• NCBI - http//www.ncbi.nlm.nih.gov
• Ensemble - http//www.ensembl.org/
• UCSC - http//genome.ucsc.edu/