Identification of Transposable Elements Using Multiple Alignments of Related Genomes

1
Identification of Transposable Elements Using
Multiple Alignments of Related Genomes

• I690 Project Presentation
• By Yin Wu

2
Transposable Elements

• Transposable Elements (TE) are the chief cause of
  gapped regions in up to 10 of currently
  sequenced genomes.
• TE causes repeated alignment gaps in multiple
  genome alignment.

3
Multiple Alignments Between Related Genomes
Consider a speciation event causing the recent
divergence of genomes S1 and S2. We expect to see
some gaps in the alignment due to small
insertions and deletions. Those long and repeated
gaps are likely to be TEs. We call these gaps
Repeated Insertion Regions (RIR). On the other
hand, RIRs are the traces of TEs. By aligning
genomes of related species, it is possible to
identify TEs.
4
Previous Work

• Anat Caspi and Lior Pachter1 compared the genomes
  of four fruit fly species.
• They located most of the (currently annotated)
  TEs in the RIR.
• They Identified new instances of TE for
  known/unknown TE families.

1Anat Caspi and Lior Pachter, Indentification of
transposable elements using multiple alignments
of related genomes.
5
Previous Work (contd)
Conserved Region
Insertion Region (gap)
Annotated TEs
6
Previous Work (method)

• Multiple alignment of homologous regions of
  related genomes to find Insertion Regions (IR)
• Local alignment of each set of IRs to find
  Repeated Insertion Regions (RIR)
• Filter and assemble RIRs.
• Compare the RIRs against the BDBP1 natural TE
  annotation set.

1http//www.fruitfly.org/p_disrupt/TE.html
7
Previous Work (limitation)

• TE may be partially aligned to random sequence
  fragments by traditional multiple alignment
  method.
• Multiple alignment methods are less tolerant to
  long insertions events than Hidden Markov Model
  (HMM).
• Therefore, pairwise HMM may report more complete
  RIR than multiple alignment does.

8
This Project V.S. Previous Work

• This Project
• Align homologous regions of each pair of genomes
  using pair-wise HMM1.
• Compare the pair-wise alignments to find
  consensus gaps (IR).
• Local alignment of each set of IRs to find
  Repeated Insertion Regions (RIR)
• Filter and assemble RIRs.
• Compare the RIRs against the BDBP natural TE
  annotation set.

• Prvious Work
• Multiple alignment of homologous regions of
  related genomes to find Insertion Regions (IR)
• Local alignment of each set of IRs to find
  Repeated Insertion Regions (RIR)
• Filter and assemble RIRs.
• Compare the RIRs against the BDBP1 natural TE
  annotation set.

1Provided by Dr. Haixu Tang
9
Input and Output

• Input
• Aligned syntenic regions of the genomes of four
  species of drosophila.
• BDBP natural TE annotation set.
• Output
• RIR of the genomes
• Data Analysis
• BDBP TE coverage of the RIR set. (i.e. how much
  percent of the BDBP TE are covered by the RIR
  set.)

10
Method

• Align homologous regions of each pair of genomes
  using pair-wise HMM.
• Compare the pair-wise alignments to find
  consensus gaps (IR).
• Local alignment of each set of IRs to find
  Repeated Insertion Regions (RIR)
• Filter and assemble RIRs.
• Compare the RIRs against the BDBP natural TE
  annotation set.

11
Filter and assemble RIRs (some details)
Micro-satellite (NOT TE)
Tandem Repeats
Nested Repeats
Concatenated Repeats
12
Filter and assemble RIRs (contd)

• Micro-satellite regions Short (lt20 bp) repeats
  with close and sequential hits to self.
• Tandem repeats Long (gt30 bp) repeats which
  sequentially align to both self and to
  subcomponents in other IRs.

13
Filter and assemble RIRs (contd)

• Nested repeats Long non-overlapping (gt30 bp)
  that sequentially align to other IRs, where there
  is no intersection between the set of IRs to
  which each subcomponent aligned.
• Concatenated repeats IRs within a certain
  genomic distance (lt700 bp) that align
  sequentially to other insertion regions.

14
Thanks to

• Dr. Haixu Tang