Ab Initio Gene Discovery

1
Ab Initio Gene Discovery

• Alberto Riva
• MGM / UFGI
• GMS6181

2
Genome annotation

• Sequencing a genome is only the first step.
  Annotation turns a genome sequence into a source
  of useful information.
• The most basic and (relatively speaking) easiest
  annotation step is the identification of
  protein-coding genes.

3

• GAGGAACCGAGAGGCTGAGACTAACCCAGAAACATCCAATTCTCAAACT
  GAAGCTCGCACTCTCGCCTCCAGCATGAAAGTCTCTGCCGCCCTTCTGTG
  CCTGCTGCTCATAGCAGCCACCTTCATTCCCCAAGGGCTCGCTCAGCCAG
  GTAAGGCCCCCTCTTCTTCTCCTTGAACCACATTGTCTTCTCTCTGAGTT
  ATCATGGACCATCCAAGCAGACGTGGTACCCACAGTCTTGCTTTAACGCT
  ACTTTTCCAAGATAAGGTGACTCAGAAAAGGACAAGGGGTGAGCCCAACC
  ACACAGCTGCTGCTCGGCAGAGCCTGAACTAGAATTCCAGCTGTGAACCC
  CAAATCCAGCTCCTTCCAGGATTCCAGCTCTGGGAACACACTCAGCGCAG
  TTACTCCCCCAGCTGCTTCCAGCAGAGTTTGGGGATCAGGGTAATCAAAG
  AGAGGGTGGGTGTGTAGGCTGTTTCCAGACACGCTGGAGACCCAGAATCT
  GGTCTGTGCTTCATTCACCTTAGCTTCCAGAGACGGTGACTCTGCAGAGG
  TAATGAGTATCAGGGAAACTCATGACCAGGCATAGCCTATTCAGAGTCTA
  AAAGGAGGCTCATAGTGGGGCTCCCCAGCTGATCTTCCCTGGTGCTGATC
  ATCTGGATTATTGGTCCGTCTTAATGACACTTGTAGGCATTATCTAGCTT
  TAACAGCTCCTCCTTCTCTCTGTCCATTATCAATGTTATATACCCATTTT
  ACAGCATAGGAAACTGAGTCATTGGGTCAAAGATCACATTCTAGCTCTGA
  GGTATAGGCAGAAGCACTGGGATTTAATGAGCTCTTTGTCTTCTCCTGCC
  TGCCTTTTGCTTTTTCCTCATGACTCTTTTCTGCTCTTAAGATCAGAATA
  ATCCAGTTCATCCTAAAATGCTTTTTCTTTGTGGTTTATTTTCCAGATGC
  AATCAATGCCCCAGTCACCTGCTGTTATAACTTCACCAATAGGAAGATCT
  CAGTGCAGAGGCTCGCGAGCTATAGAAGAATCACCAGCAGCAAGTGTCCC
  AAAGAAGCTGTGATGTGAGTTCAGCACACCAACCTTCCCTGGCCTGAAGT
  TCTTCCTTGTGGAGCAAGGGACAAGCCTCATAAACCTAGAGTCAGAGAGT
  GCACTATTTAACTTAATGTACAAAGGTTCCCAATGGGAAAACTGAGGCAC
  CAAGGGAAAAAGTGAACCCCAACATCACTCTCCACCTGGGTGCCTATTCA
  GAACACCCCAATTTCTTTAGCTTGAAGTCAGGATGGCTCCACCTGGACAC
  CTATAGGAGCAGTTTGCCCTGGGTTCCCTCCTTCCACCTGCGTTCCTCCT
  CTAGCTCCCATGGCAGCCCTTTGGTGCAGAATGGGCTGCACTTCTAGACC
  AAAACTGCAAAGGAACTTCATCTAACTCTGTCCTCCCTCCCCACAGCTTC
  AAGACCATTGTGGCCAAGGAGATCTGTGCTGACCCCAAGCAGAAGTGGGT
  TCAGGATTCCATGGACCACCTGGACAAGCAAACCCAAACTCCGAAGACTT
  GAACACTCACTCCACAACCCAAGAATCTGCAGCTAACTTATTTTCCCCTA
  GCTTTCCCCAGACACCCTGTTTTATTTTATTATAATGAATTTTGTTTGTT
  GATGTGAAACATTATGCCTTAAGTAATGTTAATTCTTATTTAAGTTATTG
  ATGTTTTAAGTTTATCTTTCATGGTACTAGTGTTTTTTAGATACAGAGAC
  TTGGGGAAATTGCTTTTCCTCTTGAACCACAGTTCTACCCCTGGGATGTT
  TTGAGGGTCTTTGCAAGAATCATTAATACAAAGAATTTTTTTTAACATTC
  CAATGCATTGCTAAAATATTATTGTGGAAATGAATATTTTGTAACTATTA
  CACCAAATAAATATATTTTTGTAC

Is there a gene in this sequence? What is its
structure? What does it do?
4
(No Transcript)
5
Some numbers

• Human genes in RefSeq 25,041
• Average number of exons 10.35
• Average exon length 270bp
• Longest exon 21,693bp(exon 81 of 83 in gene
  MUC16, transcript length is 132,498bp).

6

• GAGGAACCGAGAGGCTGAGACTAACCCAGAAACATCCAATTCTCAAACT
  GAAGCTCGCACTCTCGCCTCCAGCATGAAAGTCTCTGCCGCCCTTCTGTG
  CCTGCTGCTCATAGCAGCCACCTTCATTCCCCAAGGGCTCGCTCAGCCAG
  GTAAGGCCCCCTCTTCTTCTCCTTGAACCACATTGTCTTCTCTCTGAGTT
  ATCATGGACCATCCAAGCAGACGTGGTACCCACAGTCTTGCTTTAACGCT
  ACTTTTCCAAGATAAGGTGACTCAGAAAAGGACAAGGGGTGAGCCCAACC
  ACACAGCTGCTGCTCGGCAGAGCCTGAACTAGAATTCCAGCTGTGAACCC
  CAAATCCAGCTCCTTCCAGGATTCCAGCTCTGGGAACACACTCAGCGCAG
  TTACTCCCCCAGCTGCTTCCAGCAGAGTTTGGGGATCAGGGTAATCAAAG
  AGAGGGTGGGTGTGTAGGCTGTTTCCAGACACGCTGGAGACCCAGAATCT
  GGTCTGTGCTTCATTCACCTTAGCTTCCAGAGACGGTGACTCTGCAGAGG
  TAATGAGTATCAGGGAAACTCATGACCAGGCATAGCCTATTCAGAGTCTA
  AAAGGAGGCTCATAGTGGGGCTCCCCAGCTGATCTTCCCTGGTGCTGATC
  ATCTGGATTATTGGTCCGTCTTAATGACACTTGTAGGCATTATCTAGCTT
  TAACAGCTCCTCCTTCTCTCTGTCCATTATCAATGTTATATACCCATTTT
  ACAGCATAGGAAACTGAGTCATTGGGTCAAAGATCACATTCTAGCTCTGA
  GGTATAGGCAGAAGCACTGGGATTTAATGAGCTCTTTGTCTTCTCCTGCC
  TGCCTTTTGCTTTTTCCTCATGACTCTTTTCTGCTCTTAAGATCAGAATA
  ATCCAGTTCATCCTAAAATGCTTTTTCTTTGTGGTTTATTTTCCAGATGC
  AATCAATGCCCCAGTCACCTGCTGTTATAACTTCACCAATAGGAAGATCT
  CAGTGCAGAGGCTCGCGAGCTATAGAAGAATCACCAGCAGCAAGTGTCCC
  AAAGAAGCTGTGATGTGAGTTCAGCACACCAACCTTCCCTGGCCTGAAGT
  TCTTCCTTGTGGAGCAAGGGACAAGCCTCATAAACCTAGAGTCAGAGAGT
  GCACTATTTAACTTAATGTACAAAGGTTCCCAATGGGAAAACTGAGGCAC
  CAAGGGAAAAAGTGAACCCCAACATCACTCTCCACCTGGGTGCCTATTCA
  GAACACCCCAATTTCTTTAGCTTGAAGTCAGGATGGCTCCACCTGGACAC
  CTATAGGAGCAGTTTGCCCTGGGTTCCCTCCTTCCACCTGCGTTCCTCCT
  CTAGCTCCCATGGCAGCCCTTTGGTGCAGAATGGGCTGCACTTCTAGACC
  AAAACTGCAAAGGAACTTCATCTAACTCTGTCCTCCCTCCCCACAGCTTC
  AAGACCATTGTGGCCAAGGAGATCTGTGCTGACCCCAAGCAGAAGTGGGT
  TCAGGATTCCATGGACCACCTGGACAAGCAAACCCAAACTCCGAAGACTT
  GAACACTCACTCCACAACCCAAGAATCTGCAGCTAACTTATTTTCCCCTA
  GCTTTCCCCAGACACCCTGTTTTATTTTATTATAATGAATTTTGTTTGTT
  GATGTGAAACATTATGCCTTAAGTAATGTTAATTCTTATTTAAGTTATTG
  ATGTTTTAAGTTTATCTTTCATGGTACTAGTGTTTTTTAGATACAGAGAC
  TTGGGGAAATTGCTTTTCCTCTTGAACCACAGTTCTACCCCTGGGATGTT
  TTGAGGGTCTTTGCAAGAATCATTAATACAAAGAATTTTTTTTAACATTC
  CAATGCATTGCTAAAATATTATTGTGGAAATGAATATTTTGTAACTATTA
  CACCAAATAAATATATTTTTGTAC

7
A few hints

• The sequence is a human gene transcript, on the
  forward strand
• All splice junctions are canonical (GT-AG)
• Will an approach based on simple string search
  work?

8

• GAGGAACCGAGAGGCTGAGACTAACCCAGAAACATCCAATTCTCAAACT
  GAAGCTCGCACTCTCGCCTCCAGCATGAAAGTCTCTGCCGCCCTTCTGTG
  CCTGCTGCTCATAGCAGCCACCTTCATTCCCCAAGGGCTCGCTCAGCCAG
  GTAAGGCCCCCTCTTCTTCTCCTTGAACCACATTGTCTTCTCTCTGAGTT
  ATCATGGACCATCCAAGCAGACGTGGTACCCACAGTCTTGCTTTAACGCT
  ACTTTTCCAAGATAAGGTGACTCAGAAAAGGACAAGGGGTGAGCCCAACC
  ACACAGCTGCTGCTCGGCAGAGCCTGAACTAGAATTCCAGCTGTGAACCC
  CAAATCCAGCTCCTTCCAGGATTCCAGCTCTGGGAACACACTCAGCGCAG
  TTACTCCCCCAGCTGCTTCCAGCAGAGTTTGGGGATCAGGGTAATCAAAG
  AGAGGGTGGGTGTGTAGGCTGTTTCCAGACACGCTGGAGACCCAGAATCT
  GGTCTGTGCTTCATTCACCTTAGCTTCCAGAGACGGTGACTCTGCAGAGG
  TAATGAGTATCAGGGAAACTCATGACCAGGCATAGCCTATTCAGAGTCTA
  AAAGGAGGCTCATAGTGGGGCTCCCCAGCTGATCTTCCCTGGTGCTGATC
  ATCTGGATTATTGGTCCGTCTTAATGACACTTGTAGGCATTATCTAGCTT
  TAACAGCTCCTCCTTCTCTCTGTCCATTATCAATGTTATATACCCATTTT
  ACAGCATAGGAAACTGAGTCATTGGGTCAAAGATCACATTCTAGCTCTGA
  GGTATAGGCAGAAGCACTGGGATTTAATGAGCTCTTTGTCTTCTCCTGCC
  TGCCTTTTGCTTTTTCCTCATGACTCTTTTCTGCTCTTAAGATCAGAATA
  ATCCAGTTCATCCTAAAATGCTTTTTCTTTGTGGTTTATTTTCCAGATGC
  AATCAATGCCCCAGTCACCTGCTGTTATAACTTCACCAATAGGAAGATCT
  CAGTGCAGAGGCTCGCGAGCTATAGAAGAATCACCAGCAGCAAGTGTCCC
  AAAGAAGCTGTGATGTGAGTTCAGCACACCAACCTTCCCTGGCCTGAAGT
  TCTTCCTTGTGGAGCAAGGGACAAGCCTCATAAACCTAGAGTCAGAGAGT
  GCACTATTTAACTTAATGTACAAAGGTTCCCAATGGGAAAACTGAGGCAC
  CAAGGGAAAAAGTGAACCCCAACATCACTCTCCACCTGGGTGCCTATTCA
  GAACACCCCAATTTCTTTAGCTTGAAGTCAGGATGGCTCCACCTGGACAC
  CTATAGGAGCAGTTTGCCCTGGGTTCCCTCCTTCCACCTGCGTTCCTCCT
  CTAGCTCCCATGGCAGCCCTTTGGTGCAGAATGGGCTGCACTTCTAGACC
  AAAACTGCAAAGGAACTTCATCTAACTCTGTCCTCCCTCCCCACAGCTTC
  AAGACCATTGTGGCCAAGGAGATCTGTGCTGACCCCAAGCAGAAGTGGGT
  TCAGGATTCCATGGACCACCTGGACAAGCAAACCCAAACTCCGAAGACTT
  GAACACTCACTCCACAACCCAAGAATCTGCAGCTAACTTATTTTCCCCTA
  GCTTTCCCCAGACACCCTGTTTTATTTTATTATAATGAATTTTGTTTGTT
  GATGTGAAACATTATGCCTTAAGTAATGTTAATTCTTATTTAAGTTATTG
  ATGTTTTAAGTTTATCTTTCATGGTACTAGTGTTTTTTAGATACAGAGAC
  TTGGGGAAATTGCTTTTCCTCTTGAACCACAGTTCTACCCCTGGGATGTT
  TTGAGGGTCTTTGCAAGAATCATTAATACAAAGAATTTTTTTTAACATTC
  CAATGCATTGCTAAAATATTATTGTGGAAATGAATATTTTGTAACTATTA
  CACCAAATAAATATATTTTTGTAC

9

• GAGGAACCGAGAGGCTGAGACTAACCCAGAAACATCCAATTCTCAAACT
  GAAGCTCGCACTCTCGCCTCCAGCATGAAAGTCTCTGCCGCCCTTCTGTG
  CCTGCTGCTCATAGCAGCCACCTTCATTCCCCAAGGGCTCGCTCAGCCAG
  GTAAGGCCCCCTCTTCTTCTCCTTGAACCACATTGTCTTCTCTCTGAGTT
  ATCATGGACCATCCAAGCAGACGTGGTACCCACAGTCTTGCTTTAACGCT
  ACTTTTCCAAGATAAGGTGACTCAGAAAAGGACAAGGGGTGAGCCCAACC
  ACACAGCTGCTGCTCGGCAGAGCCTGAACTAGAATTCCAGCTGTGAACCC
  CAAATCCAGCTCCTTCCAGGATTCCAGCTCTGGGAACACACTCAGCGCAG
  TTACTCCCCCAGCTGCTTCCAGCAGAGTTTGGGGATCAGGGTAATCAAAG
  AGAGGGTGGGTGTGTAGGCTGTTTCCAGACACGCTGGAGACCCAGAATCT
  GGTCTGTGCTTCATTCACCTTAGCTTCCAGAGACGGTGACTCTGCAGAGG
  TAATGAGTATCAGGGAAACTCATGACCAGGCATAGCCTATTCAGAGTCTA
  AAAGGAGGCTCATAGTGGGGCTCCCCAGCTGATCTTCCCTGGTGCTGATC
  ATCTGGATTATTGGTCCGTCTTAATGACACTTGTAGGCATTATCTAGCTT
  TAACAGCTCCTCCTTCTCTCTGTCCATTATCAATGTTATATACCCATTTT
  ACAGCATAGGAAACTGAGTCATTGGGTCAAAGATCACATTCTAGCTCTGA
  GGTATAGGCAGAAGCACTGGGATTTAATGAGCTCTTTGTCTTCTCCTGCC
  TGCCTTTTGCTTTTTCCTCATGACTCTTTTCTGCTCTTAAGATCAGAATA
  ATCCAGTTCATCCTAAAATGCTTTTTCTTTGTGGTTTATTTTCCAGATGC
  AATCAATGCCCCAGTCACCTGCTGTTATAACTTCACCAATAGGAAGATCT
  CAGTGCAGAGGCTCGCGAGCTATAGAAGAATCACCAGCAGCAAGTGTCCC
  AAAGAAGCTGTGATGTGAGTTCAGCACACCAACCTTCCCTGGCCTGAAGT
  TCTTCCTTGTGGAGCAAGGGACAAGCCTCATAAACCTAGAGTCAGAGAGT
  GCACTATTTAACTTAATGTACAAAGGTTCCCAATGGGAAAACTGAGGCAC
  CAAGGGAAAAAGTGAACCCCAACATCACTCTCCACCTGGGTGCCTATTCA
  GAACACCCCAATTTCTTTAGCTTGAAGTCAGGATGGCTCCACCTGGACAC
  CTATAGGAGCAGTTTGCCCTGGGTTCCCTCCTTCCACCTGCGTTCCTCCT
  CTAGCTCCCATGGCAGCCCTTTGGTGCAGAATGGGCTGCACTTCTAGACC
  AAAACTGCAAAGGAACTTCATCTAACTCTGTCCTCCCTCCCCACAGCTTC
  AAGACCATTGTGGCCAAGGAGATCTGTGCTGACCCCAAGCAGAAGTGGGT
  TCAGGATTCCATGGACCACCTGGACAAGCAAACCCAAACTCCGAAGACTT
  GAACACTCACTCCACAACCCAAGAATCTGCAGCTAACTTATTTTCCCCTA
  GCTTTCCCCAGACACCCTGTTTTATTTTATTATAATGAATTTTGTTTGTT
  GATGTGAAACATTATGCCTTAAGTAATGTTAATTCTTATTTAAGTTATTG
  ATGTTTTAAGTTTATCTTTCATGGTACTAGTGTTTTTTAGATACAGAGAC
  TTGGGGAAATTGCTTTTCCTCTTGAACCACAGTTCTACCCCTGGGATGTT
  TTGAGGGTCTTTGCAAGAATCATTAATACAAAGAATTTTTTTTAACATTC
  CAATGCATTGCTAAAATATTATTGTGGAAATGAATATTTTGTAACTATTA
  CACCAAATAAATATATTTTTGTAC

10

• GAGGAACCGAGAGGCTGAGACTAACCCAGAAACATCCAATTCTCAAACT
  GAAGCTCGCACTCTCGCCTCCAGCATGAAAGTCTCTGCCGCCCTTCTGTG
  CCTGCTGCTCATAGCAGCCACCTTCATTCCCCAAGGGCTCGCTCAGCCAG
  GTAAGGCCCCCTCTTCTTCTCCTTGAACCACATTGTCTTCTCTCTGAGTT
  ATCATGGACCATCCAAGCAGACGTGGTACCCACAGTCTTGCTTTAACGCT
  ACTTTTCCAAGATAAGGTGACTCAGAAAAGGACAAGGGGTGAGCCCAACC
  ACACAGCTGCTGCTCGGCAGAGCCTGAACTAGAATTCCAGCTGTGAACCC
  CAAATCCAGCTCCTTCCAGGATTCCAGCTCTGGGAACACACTCAGCGCAG
  TTACTCCCCCAGCTGCTTCCAGCAGAGTTTGGGGATCAGGGTAATCAAAG
  AGAGGGTGGGTGTGTAGGCTGTTTCCAGACACGCTGGAGACCCAGAATCT
  GGTCTGTGCTTCATTCACCTTAGCTTCCAGAGACGGTGACTCTGCAGAGG
  TAATGAGTATCAGGGAAACTCATGACCAGGCATAGCCTATTCAGAGTCTA
  AAAGGAGGCTCATAGTGGGGCTCCCCAGCTGATCTTCCCTGGTGCTGATC
  ATCTGGATTATTGGTCCGTCTTAATGACACTTGTAGGCATTATCTAGCTT
  TAACAGCTCCTCCTTCTCTCTGTCCATTATCAATGTTATATACCCATTTT
  ACAGCATAGGAAACTGAGTCATTGGGTCAAAGATCACATTCTAGCTCTGA
  GGTATAGGCAGAAGCACTGGGATTTAATGAGCTCTTTGTCTTCTCCTGCC
  TGCCTTTTGCTTTTTCCTCATGACTCTTTTCTGCTCTTAAGATCAGAATA
  ATCCAGTTCATCCTAAAATGCTTTTTCTTTGTGGTTTATTTTCCAGATGC
  AATCAATGCCCCAGTCACCTGCTGTTATAACTTCACCAATAGGAAGATCT
  CAGTGCAGAGGCTCGCGAGCTATAGAAGAATCACCAGCAGCAAGTGTCCC
  AAAGAAGCTGTGATGTGAGTTCAGCACACCAACCTTCCCTGGCCTGAAGT
  TCTTCCTTGTGGAGCAAGGGACAAGCCTCATAAACCTAGAGTCAGAGAGT
  GCACTATTTAACTTAATGTACAAAGGTTCCCAATGGGAAAACTGAGGCAC
  CAAGGGAAAAAGTGAACCCCAACATCACTCTCCACCTGGGTGCCTATTCA
  GAACACCCCAATTTCTTTAGCTTGAAGTCAGGATGGCTCCACCTGGACAC
  CTATAGGAGCAGTTTGCCCTGGGTTCCCTCCTTCCACCTGCGTTCCTCCT
  CTAGCTCCCATGGCAGCCCTTTGGTGCAGAATGGGCTGCACTTCTAGACC
  AAAACTGCAAAGGAACTTCATCTAACTCTGTCCTCCCTCCCCACAGCTTC
  AAGACCATTGTGGCCAAGGAGATCTGTGCTGACCCCAAGCAGAAGTGGGT
  TCAGGATTCCATGGACCACCTGGACAAGCAAACCCAAACTCCGAAGACTT
  GAACACTCACTCCACAACCCAAGAATCTGCAGCTAACTTATTTTCCCCTA
  GCTTTCCCCAGACACCCTGTTTTATTTTATTATAATGAATTTTGTTTGTT
  GATGTGAAACATTATGCCTTAAGTAATGTTAATTCTTATTTAAGTTATTG
  ATGTTTTAAGTTTATCTTTCATGGTACTAGTGTTTTTTAGATACAGAGAC
  TTGGGGAAATTGCTTTTCCTCTTGAACCACAGTTCTACCCCTGGGATGTT
  TTGAGGGTCTTTGCAAGAATCATTAATACAAAGAATTTTTTTTAACATTC
  CAATGCATTGCTAAAATATTATTGTGGAAATGAATATTTTGTAACTATTA
  CACCAAATAAATATATTTTTGTAC

11

• GAGGAACCGAGAGGCTGAGACTAACCCAGAAACATCCAATTCTCAAACT
  GAAGCTCGCACTCTCGCCTCCAGCATGAAAGTCTCTGCCGCCCTTCTGTG
  CCTGCTGCTCATAGCAGCCACCTTCATTCCCCAAGGGCTCGCTCAGCCAG
  GTAAGGCCCCCTCTTCTTCTCCTTGAACCACATTGTCTTCTCTCTGAGTT
  ATCATGGACCATCCAAGCAGACGTGGTACCCACAGTCTTGCTTTAACGCT
  ACTTTTCCAAGATAAGGTGACTCAGAAAAGGACAAGGGGTGAGCCCAACC
  ACACAGCTGCTGCTCGGCAGAGCCTGAACTAGAATTCCAGCTGTGAACCC
  CAAATCCAGCTCCTTCCAGGATTCCAGCTCTGGGAACACACTCAGCGCAG
  TTACTCCCCCAGCTGCTTCCAGCAGAGTTTGGGGATCAGGGTAATCAAAG
  AGAGGGTGGGTGTGTAGGCTGTTTCCAGACACGCTGGAGACCCAGAATCT
  GGTCTGTGCTTCATTCACCTTAGCTTCCAGAGACGGTGACTCTGCAGAGG
  TAATGAGTATCAGGGAAACTCATGACCAGGCATAGCCTATTCAGAGTCTA
  AAAGGAGGCTCATAGTGGGGCTCCCCAGCTGATCTTCCCTGGTGCTGATC
  ATCTGGATTATTGGTCCGTCTTAATGACACTTGTAGGCATTATCTAGCTT
  TAACAGCTCCTCCTTCTCTCTGTCCATTATCAATGTTATATACCCATTTT
  ACAGCATAGGAAACTGAGTCATTGGGTCAAAGATCACATTCTAGCTCTGA
  GGTATAGGCAGAAGCACTGGGATTTAATGAGCTCTTTGTCTTCTCCTGCC
  TGCCTTTTGCTTTTTCCTCATGACTCTTTTCTGCTCTTAAGATCAGAATA
  ATCCAGTTCATCCTAAAATGCTTTTTCTTTGTGGTTTATTTTCCAGATGC
  AATCAATGCCCCAGTCACCTGCTGTTATAACTTCACCAATAGGAAGATCT
  CAGTGCAGAGGCTCGCGAGCTATAGAAGAATCACCAGCAGCAAGTGTCCC
  AAAGAAGCTGTGATGTGAGTTCAGCACACCAACCTTCCCTGGCCTGAAGT
  TCTTCCTTGTGGAGCAAGGGACAAGCCTCATAAACCTAGAGTCAGAGAGT
  GCACTATTTAACTTAATGTACAAAGGTTCCCAATGGGAAAACTGAGGCAC
  CAAGGGAAAAAGTGAACCCCAACATCACTCTCCACCTGGGTGCCTATTCA
  GAACACCCCAATTTCTTTAGCTTGAAGTCAGGATGGCTCCACCTGGACAC
  CTATAGGAGCAGTTTGCCCTGGGTTCCCTCCTTCCACCTGCGTTCCTCCT
  CTAGCTCCCATGGCAGCCCTTTGGTGCAGAATGGGCTGCACTTCTAGACC
  AAAACTGCAAAGGAACTTCATCTAACTCTGTCCTCCCTCCCCACAGCTTC
  AAGACCATTGTGGCCAAGGAGATCTGTGCTGACCCCAAGCAGAAGTGGGT
  TCAGGATTCCATGGACCACCTGGACAAGCAAACCCAAACTCCGAAGACTT
  GAACACTCACTCCACAACCCAAGAATCTGCAGCTAACTTATTTTCCCCTA
  GCTTTCCCCAGACACCCTGTTTTATTTTATTATAATGAATTTTGTTTGTT
  GATGTGAAACATTATGCCTTAAGTAATGTTAATTCTTATTTAAGTTATTG
  ATGTTTTAAGTTTATCTTTCATGGTACTAGTGTTTTTTAGATACAGAGAC
  TTGGGGAAATTGCTTTTCCTCTTGAACCACAGTTCTACCCCTGGGATGTT
  TTGAGGGTCTTTGCAAGAATCATTAATACAAAGAATTTTTTTTAACATTC
  CAATGCATTGCTAAAATATTATTGTGGAAATGAATATTTTGTAACTATTA
  CACCAAATAAATATATTTTTGTAC

12
So the problem is simply

• Find one or more GT-AG pairs such that removing
  all the nucleotides between them, what remains
  (between one of the ATGs and one of the stop
  codons) can be correctly translated into an amino
  acid sequence.
• (Gene prediction is much, much easier if we know
  the cDNA or protein sequence becomes a sequence
  alignment problem).

13
ORF finding

1. Translate DNA sequence in all three (or six)
   frames
2. Look for ORFs, stretches with no stop codons
   longer than what would be expected at random
3. Look for coding exons within the ORFs.

14
ORF finding

• Looking for ORFs
• 3 stop codons out of 64 1 every 21.3 amino
  acids. Average ORF length is 64bp.
• Unfortunately, many exons are shorter. A 51bp ORF
  could be an exon, or a false positive.

15
ORF finding

• Looking for exons within ORFs
• ORFs extend before and after exons. Identifying
  the exact start and end site is hard. Only works
  for coding exons.
• Exons can be in any one of the three frames.
  Alternative splicing, mutually exclusive exons
  add complexity.

16
ORF finding - example

• Our sequence contains 85 stop codons
• 27 in frame 1
• 35 in frame 2
• 23 in frame 3
• Average ORF length is 65.5, as expected
• 29 ORFs out of 82 (35) are longer.

17
ORFs in frame 2 GAGGAACCGAGAGGCTGAGACTAACC
CAGAAACATCCAATTCTCAAACTGAAGCTCGCACTCTCGCCTCCAGCATG
AAAGTCTCTGCCGCCCTTCTGTGCCTGCTGCTCATAGCAGCCACCTTCAT
TCCCCAAGGGCTCGCTCAGCCAGGTAAGGCCCCCTCTTCTTCTCCTTGAA
CCACATTGTCTTCTCTCTGAGTTATCATGGACCATCCAAGCAGACGTGGT
ACCCACAGTCTTGCTTTAACGCTACTTTTCCAAGATAAGGTGACTCAGAA
AAGGACAAGGGGTGAGCCCAACCACACAGCTGCTGCTCGGCAGAGCCTGA
ACTAGAATTCCAGCTGTGAACCCCAAATCCAGCTCCTTCCAGGATTCCAG
CTCTGGGAACACACTCAGCGCAGTTACTCCCCCAGCTGCTTCCAGCAGAG
TTTGGGGATCAGGGTAATCAAAGAGAGGGTGGGTGTGTAGGCTGTTTCCA
GACACGCTGGAGACCCAGAATCTGGTCTGTGCTTCATTCACCTTAGCTTC
CAGAGACGGTGACTCTGCAGAGGTAATGAGTATCAGGGAAACTCATGACC
AGGCATAGCCTATTCAGAGTCTAAAAGGAGGCTCATAGTGGGGCTCCCCA
GCTGATCTTCCCTGGTGCTGATCATCTGGATTATTGGTCCGTCTTAATGA
CACTTGTAGGCATTATCTAGCTTTAACAGCTCCTCCTTCTCTCTGTCCAT
TATCAATGTTATATACCCATTTTACAGCATAGGAAACTGAGTCATTGGGT
CAAAGATCACATTCTAGCTCTGAGGTATAGGCAGAAGCACTGGGATTTAA
TGAGCTCTTTGTCTTCTCCTGCCTGCCTTTTGCTTTTTCCTCATGACTCT
TTTCTGCTCTTAAGATCAGAATAATCCAGTTCATCCTAAAATGCTTTTTC
TTTGTGGTTTATTTTCCAGATGCAATCAATGCCCCAGTCACCTGCTGTTA
TAACTTCACCAATAGGAAGATCTCAGTGCAGAGGCTCGCGAGCTATAGAA
GAATCACCAGCAGCAAGTGTCCCAAAGAAGCTGTGATGTGAGTTCAGCAC
ACCAACCTTCCCTGGCCTGAAGTTCTTCCTTGTGGAGCAAGGGACAAGCC
TCATAAACCTAGAGTCAGAGAGTGCACTATTTAACTTAATGTACAAAGGT
TCCCAATGGGAAAACTGAGGCACCAAGGGAAAAAGTGAACCCCAACATCA
CTCTCCACCTGGGTGCCTATTCAGAACACCCCAATTTCTTTAGCTTGAAG
TCAGGATGGCTCCACCTGGACACCTATAGGAGCAGTTTGCCCTGGGTTCC
CTCCTTCCACCTGCGTTCCTCCTCTAGCTCCCATGGCAGCCCTTTGGTGC
AGAATGGGCTGCACTTCTAGACCAAAACTGCAAAGGAACTTCATCTAACT
CTGTCCTCCCTCCCCACAGCTTCAAGACCATTGTGGCCAAGGAGATCTGT
GCTGACCCCAAGCAGAAGTGGGTTCAGGATTCCATGGACCACCTGGACAA
GCAAACCCAAACTCCGAAGACTTGAACACTCACTCCACAACCCAAGAATC
TGCAGCTAACTTATTTTCCCCTAGCTTTCCCCAGACACCCTGTTTTATTT
TATTATAATGAATTTTGTTTGTTGATGTGAAACATTATGCCTTAAGTAAT
GTTAATTCTTATTTAAGTTATTGATGTTTTAAGTTTATCTTTCATGGTAC
TAGTGTTTTTTAGATACAGAGACTTGGGGAAATTGCTTTTCCTCTTGAAC
CACAGTTCTACCCCTGGGATGTTTTGAGGGTCTTTGCAAGAATCATTAAT
ACAAAGAATTTTTTTTAACATTCCAATGCATTGCTAAAATATTATTGTGG
AAATGAATATTTTGTAACTATTACACCAAATAAATATATTTTTGTAC
18
Help!

• More informative consensus sequences for splice
  sites
• AGGTRAGT
• YYYNCAGG (RA/G, YC/T).
• Information on expected length of exons and
  introns, codon usage, hexamer frequencies , CG
  content.
• ? Probabilistic methods

19
Probabilistic methods

• Probabilistic methods compute the probability
  that a feature (e.g. an exon) is present in a
  certain region of the input sequence.
• They are able to combine different forms of
  evidence into a score for each detected
  feature.
• A good example is GENSCAN, based on Hidden Markov
  Models (HMMs). http//genes.mit.edu/GENSCAN.html

20
GENSCAN model

• An HMM is a state machine, with probabilities
  describing transitions between states.
• Can take into account correlations between
  exons, to build complete and accurate models.

21
GENSCAN output

• GENSCANW output for sequence 211632
• GENSCAN 1.0 Date run 12-Jan-107 Time 211635
• Sequence Transcript 1923 bp 45.66 CG
  Isochore 2 (43 - 51 CG)
• Parameter matrix HumanIso.smat
• Predicted genes/exons
• Gn.Ex Type S .Begin ...End .Len Fr Ph I/Ac Do/T
  CodRg P.... Tscr..
• ----- ---- - ------ ------ ---- -- -- ---- ----
  ----- ----- ------
• 1.01 Init 74 149 76 1 1 69 110
  69 0.934 7.40
• 1.02 Intr 946 1063 118 2 1 105 76
  94 0.992 9.42
• 1.03 Term 1446 1551 106 0 1 105 47
  127 0.996 8.18
• 1.04 PlyA 1905 1910 6
  1.05
• Predicted peptide sequence(s)
  gtTranscriptGENSCAN_predicted_peptide_199_aa
  MKVSAALLCLLLIAATFIPQGLAQPDAINAPVTCCYNFTNRKISVQRLAS
  YRRITSSKCP KEAVIFKTIVAKEICADPKQKWVQDSMDHLDKQTQTPKT

22
gaggaaccgagaggctgagactaacccagaaacatccaattctcaaact
gaagctcgcactctcgcctccagcATGAAAGTCTCTGCCGCCCTTCTGTG
CCTGCTGCTCATAGCAGCCACCTTCATTCCCCAAGGGCTCGCTCAGCCAG
GTAAGGCCCCCTCTTCTTCTCCTTGAACCACATTGTCTTCTCTCTGAGTT
ATCATGGACCATCCAAGCAGACGTGGTACCCACAGTCTTGCTTTAACGCT
ACTTTTCCAAGATAAGGTGACTCAGAAAAGGACAAGGGGTGAGCCCAACC
ACACAGCTGCTGCTCGGCAGAGCCTGAACTAGAATTCCAGCTGTGAACCC
CAAATCCAGCTCCTTCCAGGATTCCAGCTCTGGGAACACACTCAGCGCAG
TTACTCCCCCAGCTGCTTCCAGCAGAGTTTGGGGATCAGGGTAATCAAAG
AGAGGGTGGGTGTGTAGGCTGTTTCCAGACACGCTGGAGACCCAGAATCT
GGTCTGTGCTTCATTCACCTTAGCTTCCAGAGACGGTGACTCTGCAGAGG
TAATGAGTATCAGGGAAACTCATGACCAGGCATAGCCTATTCAGAGTCTA
AAAGGAGGCTCATAGTGGGGCTCCCCAGCTGATCTTCCCTGGTGCTGATC
ATCTGGATTATTGGTCCGTCTTAATGACACTTGTAGGCATTATCTAGCTT
TAACAGCTCCTCCTTCTCTCTGTCCATTATCAATGTTATATACCCATTTT
ACAGCATAGGAAACTGAGTCATTGGGTCAAAGATCACATTCTAGCTCTGA
GGTATAGGCAGAAGCACTGGGATTTAATGAGCTCTTTGTCTTCTCCTGCC
TGCCTTTTGCTTTTTCCTCATGACTCTTTTCTGCTCTTAAGATCAGAATA
ATCCAGTTCATCCTAAAATGCTTTTTCTTTGTGGTTTATTTTCCAGATGC
AATCAATGCCCCAGTCACCTGCTGTTATAACTTCACCAATAGGAAGATCT
CAGTGCAGAGGCTCGCGAGCTATAGAAGAATCACCAGCAGCAAGTGTCCC
AAAGAAGCTGTGATGTGAGTTCAGCACACCAACCTTCCCTGGCCTGAAGT
TCTTCCTTGTGGAGCAAGGGACAAGCCTCATAAACCTAGAGTCAGAGAGT
GCACTATTTAACTTAATGTACAAAGGTTCCCAATGGGAAAACTGAGGCAC
CAAGGGAAAAAGTGAACCCCAACATCACTCTCCACCTGGGTGCCTATTCA
GAACACCCCAATTTCTTTAGCTTGAAGTCAGGATGGCTCCACCTGGACAC
CTATAGGAGCAGTTTGCCCTGGGTTCCCTCCTTCCACCTGCGTTCCTCCT
CTAGCTCCCATGGCAGCCCTTTGGTGCAGAATGGGCTGCACTTCTAGACC
AAAACTGCAAAGGAACTTCATCTAACTCTGTCCTCCCTCCCCACAGCTTC
AAGACCATTGTGGCCAAGGAGATCTGTGCTGACCCCAAGCAGAAGTGGGT
TCAGGATTCCATGGACCACCTGGACAAGCAAACCCAAACTCCGAAGACTT
GAacactcactccacaacccaagaatctgcagctaacttattttccccta
gctttccccagacaccctgttttattttattataatgaattttgtttgtt
gatgtgaaacattatgccttaagtaatgttaattcttatttaagttattg
atgttttaagtttatctttcatggtactagtgttttttagatacagagac
ttggggaaattgcttttcctcttgaaccacagttctacccctgggatgtt
ttgagggtctttgcaagaatcattaatacaaagaattttttttaacattc
caatgcattgctaaaatattattgtggaaatgaatattttgtaactatta
caccaaataaatatatttttgtac
23
What did we find?

• We discovered gene CCL2 (Small inducible cytokine
  A2 precursor), on chromosome 17.
• Next steps look at similarity with other genes,
  homologs in different organisms, protein sequence
  structure in order to infer gene function.

24
State of the art

• In real life long sequences containing many
  genes, on both strands. Overlapping genes. Genes
  in introns of other genes.
• 5 of RefSeq genes have incorrect structure
  (wrong translation).
• How could we emulate what the transcriptional/spli
  cing machinery does?

25
Future challenges

• Detecting promoters, control regions,
  transcription factor binding sites, splicing
  enhancers and silencers
• Non-coding RNAs, micro-RNAs, etc
• Insulators, boundary elements, matrix-attachment
  and scaffold-attachment regions, replication
  origins, recombination hotspots ???