A computational view of microRNAs and their targets

1
A computational view of microRNAs and their
targets

• ??? 2005/8/23

2
Introduction (1/3)

• Small non-coding RNAs called microRNAs have been
  shown to play important roles in gene regulation
• Usually 2125 nucleotides in length
• Derived from a larger precursor , 6070
  nucleotides

3
(No Transcript)
4
Introduction (2/3)

• A majority of mammalian miRNA genes overlap with
  transcription units
• Introns of protein-coding genes
• Both introns and exons of mRNA-like non-coding
  genes
• miRNAs are transcribed in parallel with these
  host genes
• Intronic miRNAs have also been identified in C.
  elegans

5
Introduction (3/3)

• In plants , miRNAs induce cleavage
• In animals , miRNAs induce translational
  repression

6
Computational identification of miRNAs (1/4)

• Discovery of the C. elegans miRNA let-7 led to
  computational approaches
• Through simple sequence homology searches ,
  orthologues of let-7 were identified in numerous
  species
• Alternative approaches used RNA folding
  prediction

7
Computational identification of miRNAs (2/4)

• MiRscan
• RNAfold or Mfold
• Predicting gt35 000 stem-loops conserved between
  C. briggsae and C. elegans
• 50 published miRNAs from these two species were
  used as a training set
• Slide a 21-nucleotide window along candidate
  stem-loops and assign a score
• Scoring

8
Computational identification of miRNAs (3/4)

• 30 additional miRNAs were identified
• C. elegans genome probably has 120 miRNA genes
• Upper bound of 255 human miRNA genes
• Some miRNAs appear to be poorly conserved between
  humans and pufferfish

9
Computational identification of miRNAs (4/4)

• In plants
• Variability in length of miRNA precursors
• Differences in GC content
• Lower sequence conservation of precursors

10
(No Transcript)
11
Computational prediction of miRNA targets (1/4)

• Based on knowledge of base pairing between the
  miRNA and the target gene
• The criteria are the complementarity between the
  3 UTRs of the potential targets and miRNAs
• Emphasis on the critical pairing at the 5 end of
  the miRNA

12
Computational prediction of miRNA targets (2/4)

• The conservation of target 3 UTR sequences in
  orthologous genes
• Free energy
• Key differences

13
Computational prediction of miRNA targets (3/4)

• TargetScanS
• The omission of both the thermodynamic stability
  of pairings and multiple target sites per UTR
• Seven-nucleotide match relax to six-nucleotide
  match
• Using this refined algorithm on four genomes
  (humans, mouse, rat and dog)
• 3 UTRs of 30 of the gene set used in the
  analysis were potential miRNA targets

14
Computational prediction of miRNA targets (4/4)

• In plants
• Near to perfect complementarity
• Animal miRNA targets appear to be more
  functionally diverse than plant
• Plant miRNAs appear to target in particular
  transcription factors involved in development

15
(No Transcript)
16
miRNA data resources

• miRNA Registry
• Rename to miRBase
• Precursor and mature sequences
• The criteria for the inclusion of miRNA sequences
  in the database are not identical across species

17

• http//web.wi.mit.edu/bartel/pub/rnalinks.html