Annotating mRNAs

1
Annotating mRNAs proteins

• Goals
• Improve melanogaster annotation by using new
  genome sequences
• Annotate new genome sequences
• Provide a model for human annotation

2
Dual genome predictorsThe state of the art
3
Right multi-genome model yields accuracy
breakthrough
Mouse, rat, chicken
GENSCAN
Mouse comparison
Many small improvements
4
Verification by RT-PCR

• Infrastructure in place
• Automated target selection primer design
• Automated analysis of results
• Shared archival databases for experiments
• Cost is low
• PCR mix sequenced directly (skip the gel)
• Supplies services 10 / RT-PCR 2 reads
• 55 of novel elegans predictions confirmed

5
We predict TSS and UTR exons in mammals
6
Traces reveal alternative splices
7
Translation start sites

• State of the art
• Guess! Longest ORF or 1st ATG or Kozak
• I am not aware of any standard methodology
• If there are experimentally confirmed starts, I
  dont know how to find them in the databases
• What can be done?
• Old fashioned protein sequencing?
• Mass spectrometry?
• What about N-terminal clipping?

8
Why buy the fly and spurn the worm?

• Better model for human
• Intron length prediction difficulty
• Species at a variety of evolutionary distances
• No pair of worms is as close as mammals
• Better system for method development
• Species at a variety of evolutionary distances