mRNA Processing

1
mRNA Processing
2
RNA splicing

• Splice sites are encoded in the sequence.
• Splice site recognition is complex and imperfect.

3
Alternative splicing

• Genes can produce multiple mRNA isoforms
• Human avg 5 mRNA/gene highly skewed
• Fewer introns/gene correlates w/ fewer
  isoforms/genes

• UCSC genome browser http//genome.ucsc.edu

4
Genes

• Molecular definition
• DNA regions that are transcribed into a single
  RNA strand, with nearby DNA regions controlling
  time and quantity of transcription
• Protein-coding genes and ncRNA genes
• Classical definition
• Whatever it is that gives rise to a heritable
  trait

5
Genome sizes

• Widely varied
• Not well correlated with organism
  complexity/sophistication
• Typical bacterium 1-10 megabases (mb)
• Typical single-celled eukaryote 10-30 mb
• Smallest plants and animals 100 mb (fruit fly,
  worm, mustard weed)
• Human 3 gb some rats gophers 5-6 gb
• Pine tree 60 g Fern is 160 gb
• Database of Genome Sizes (DOGS)

6
Protein coding gene counts

• Not so widely varied
• E. coli (bacterium)
• 4,300 1200 nt/gene (no introns)
• S. cerevisiae (yeast)
• 5,800 2100 nt/gene (almost no introns)
• C. elegans (worm)
• 20,000 5000 nt/gene (5 introns/gene)
• Human other mammals
• 21,000 143,00 nt/gene (10 introns/gene)

7
Common human repeats

• LINES Long interspersed elements
• Most common is LINE1
• 6-7 kb
• 60,000 copies in human. 15 of genome.
• SINES Short interspersed elements
• Most common is Alu
• About 300 bp
• 1 million copies. 10 of human genome.
• Low complexity
• E.g., 5-10bp tandemly repeated repeated
• Especially near centromeres telomeres.

8
Computational problems Predicting exon-intron
structures

• De novo gene (mRNA) prediction Given only a
  genome sequence, predict the exon-intron
  structures of all mRNAs it encodes
• Multi-genome de novo Given also the genome
  sequences of 1 or more related organisms
• RNA-based given also partial sequences of a
  subset of RNAs from a variety of species
• Difficult in plants animals because most
  genomic DNA does not encode proteins

9
Computational Problems Categorizing proteins by
function

• After determining the mRNA products
• Relatively easy to infer the sequences of the
  proteins they encode
• Need to infer their approximate functions
• Search database of proteins with known function
  Search Alignment Problem
• Find domains (modular components) with known
  function. E.g.,
• DNA binding, ATPase, transmembrane,

10
Non-coding RNA

• Functions
• Transfer RNAs codon-to-amino-acid adapters
• Ribosomes catalyze amino acid linkage
• Protein-RNA complex. RNA is catalytic!
• Small RNAs edit specific mRNAs, or
• prevent translation of specific mRNAs
• All transcribed from DNA but not translated
• Structure
• Shape, determined by self-pairing, is essential
• External base-pairing is usually essential, too

11
RNA

• Normally single-stranded
• Much less stable than DNA. Shorter lifetime.
• Can form complex structure by self-base-pairing

12
RNA self-base-pairing
13
3D shape of transfer RNA
14
Computational problems nc RNA

• secondary structure Given an RNA sequence,
  predict its folded form -- which bases will pair
  with one another
• RNA homology Given an RNA sequence, search db
  for sequences with the same secondary structure
• Given an RNA sequence and its 2nd-ary structure,
  search db for sequences with the same 2nd-ary
  structure

15
Transcriptional regulation

• Transcription factors (TFs)
• Proteins that bind to short (8-16 nt) DNA
  sequences, affecting transcription rate
• In smaller genomes, TF binding sites are
  typically in the promoter region, upstream of a
  genes transcription start site
• Repressors reduce the genes transcription rate,
  activators increase it
• Transcription is initiated by RNA Polymerase II
  binding to the transcription start site

16
TF binding sites

• Specificity
• Each type of TF (300 in yeast) binds to a class
  of sequences called a motif
• Computational problems involving TF binding sites
• Find functional binding sites in genomic DNA
  sequence for a TF with a known motif
• Find functional binding sites without a known
  motif
• Given sequences containing sites for a given TF
  find the sites and determine the motif

17
TFs form regulatory networks

• Glucose sensing signaling in yeast

18
Developmental gene regulation
Johnson et al. 2007, Science
19
Computational problems in systems biology

• Network identification
• Given, e.g., gene expression levels in a variety
  of conditions, or a time series after stimulus,
  infer the controlling TF network
• There are various other data sources
• It helps if you know the TF binding motifs
• Network simulation
• Given a particular gene regulation network,
  predict how it will respond to novel stimuli or
  genetic manipulation, as a function of time

20
DNA Packaging

• DNA is packed hierarchically
• The chromosome is the largest package
• Humans have 22 chrs 2 sex chrs
• Human genome 2m long 0.34nm/base
• DNA is 1 picogram (10-12g) per gigabase

Quicktime animation
21
Epigenomics

• Chromatin marks
• Chemical modification of histones
• Affects DNA accessibility
• Specific marks correlated with transcription
  initiation, elongation, and silencing
• Code is not well understood

Peterson et al 2004
22
Epigenomics

• DNA methylation
• Mostly mCG, mCNG (plants)
• Heritable through cell division
• Associated with gene silencing
• Controls mobile elements
• X-inactivation and imprinting
• mC often mutates to T, leaving few CG pairs
• Except in rarely methylated regions (e.g.
  promoters of genes essential to all cell types)