Sequence Alignments

1
Sequence Alignments

• June 2, 2009

2
Outline

• How sequences are aligned
• How alignments are scored
• Understanding BLAST

3
Reading assignment

• Required
• Chapters 3 4 Xiong
• Optional
• The BLAST help tab
• http//blast.ncbi.nlm.nih.gov/Blast.cgi?CMDWebPA
  GE_TYPEBlastDocs
• There is a link from the exercise 3 homepage

4
Sequence alignment

• Determine if two sequences are related
• Sequence assembly
• Sequence annotation
• Identify shared protein domains or motifs
• Analysis of genomes
• Phylogeny and evolution

5
Definitions

• Homologous share a common ancestor
• Cannot be measured
• Measure similarity infer homology
• Orthologs separated by speciation
• Paralogs separated by duplication

6
Sequence alignment

• Determine whether two (or more) sequences are of
  sufficient similarity such that an inference of
  homology is justified

7
How to align 2 sequences

• Choose 2 sequences
• Select an algorithm
• Scoring method that reflects degree of similarity
• Allow for gaps (insertions and deletions)
• Estimate probability that alignment occurred by
  chance

8
Dotplots visual alignment
9
(No Transcript)
10
Dotplots self alignment
Alignments show up as diagonal lines on the
plot Gaps are evidenced by vertical or
horizontal shifts
11
Can find repeats

• Align sequence to itself
• Repeats are shorter diagonals off the main
  diagonal

12
Low complexity regions

• Mucin 40 exact tandem repeats of 20 amino acids

13
Compare sequences

• HMG1 and SRY are somewhat similar to each other
• But dotplots do not give a quantitative measure
  of the degree of similarity

14
blast2seq output
15
identity and similarity

• identity percentage of aligned residues that
  are identical
• similarity percentage of aligned residues that
  have similar chemical/physical properties
• Amino acid alignments only

16
Scoring schemes

• Biologically significant way of scoring matches,
  mismatches gaps
• Nucleotide alignments
• Identity only
• Positive score for matches negative scores for
  mismatches

17
Amino acid substitution matrices

• Method to score matches and mismatches
• Based on observed frequencies of amino acid
  distributions and substitutions
• Must model conservative nature of substitutions
• Implicitly represent evolutionary patterns
• Scores are based in Information Theory

18
Scoring amino acid substitutions

• Amino acids are NOT distributed evenly
• Amino acids share similarity based on chemical
  and physical properties

19
(No Transcript)
20
PAM scoring matrices

• Margaret O. Dayhoff developed first protein
  substitution matrices based on observed
  frequencies of amino acids as well as observed
  substitutions in aligned proteins (1978)
• PAM Point Accepted Mutations
• Observed variations groups of sequences 85
  similar
• estimated substitutions in a group of evolving
  proteins
• represent substitutions that do not significantly
  alter protein structure/function, so accepted
  by natural selection

21
PAM Scoring Matrices

• Based on mutational model of evolution
• Assume changes occur independently
• Changes are a prediction of first changes that
  occur as proteins diverge from common ancestor
• Matrices for more distantly related protein
  sequences extrapolated from short-term changes
• All amino acids positions in related sequences
  were scored

22
PAM Scoring matrices
S score for amino acid pairing in the alignment
qij is the observed pairing frequency of amino
acids i and j.
pi and pj are the expected frequencies for amino
acids i and j.
23
PAM 250 Matrix
24
PAM 250 Matrix
25
BLOSUM matrices

• BLOSUM matrices are based on local alignments
• BLOSUM BLOcks SUbstitution Matrix
• Sequences within segments clustered into blocks
  based on identity
• Contributions of the sequences within a cluster
  were averaged.
• BLOSUM62 is a matrix calculated from comparisons
  of sequences lt 62 identical.
• BLOSUM40 PAM250

26
BLOSUM matrices

• All BLOSUM matrices are based on observed
  alignments they are not extrapolated from
  comparisons of closely related proteins.
• The BLOCKS database contains thousands of groups
  of multiple sequence alignments.
• BLOSUM62 is the default matrix in BLAST 2.0.

27
BLOSUM62 Matrix
28
BLOSUM62 Matrix
29
BLOSUM90
More positive more negative than BLOSUM62
30
Choosing a matrix
31
Gaps

• Insertions can lead to gaps of varying lengths
• Use 2 gap penalties
• higher penalty for opening a gap
• lower penalty for extension of a gap

32
BLAST
Calculate statistical significance of matches
Build word list from query sequence
Find hits in database sequence
Extend the hits to form HSPs

• Build a list of words from query sequence
• (3 for proteins, 11 for DNA)
• Evaluate each word for match using scoring matrix
  and discard all below threshold
• Generally 50 matches per word
• T value is threshold determines sensitivity and
  speed of search

33
Query sequence
PSATPVLICWAAG
Word list
PSA ATP VLI CWA
Threshold score (T)
11
Matches to PSA Score
PSA 15 PST 9 PDA 11 WSA
4
34
BLAST
Calculate statistical significance of matches
Build word list from query sequence
Extend the hits to form HSPs
Find hits in database sequence

• Find match for each word in database
• Database is indexed so all possible words in all
  sequences is known
• This search is very fast (500K words/sec)
• Matches gt T are used as seed for alignments

35
BLAST
Calculate statistical significance of matches
Build word list from query sequence
Find hits in database sequence
Extend the hits to form HSPs

• Extend alignment from each word in both
  directions so long as score increases
• These alignments are the HSPs
• Keep HSPs if score is above a given threshold

36
Extending the hit
Score of new alignment
Score of previous alignment (A)
Score of new aligned pair


(1)
P S A C P S A C 24
p S A P S A 15
C C 9


Score of new aligned pair
(2)
Score of alignment (C)
Score of previous alignment (B)


P S A C Y P S A C Y 31
P S A C P S A C 24
Y Y 7


(3)
Repeat adding aligned pairs until score goes down
or reach end of sequence.
37
BLAST
Combine HSPs into a gapped alignment
Build word list from query sequence
Find hits in database sequence
Extend the hits to form HSPs

• Highest scoring HSPs extended in both directions
  using dynamic programming
• Continues as long as score gt threshold

38
Positives 200/310 (64)
Identities 135/310 (43)
Expect 2e-73
39
BLAST statistics

• BLASTN match with E-value .004

atgctctggccacggcacttgcggatcccagggtgatctgtgcacctgcg
ata
atgctatggccacgggtcttgtggatccca---t
gatgtgtgcacctgcgata
How is this calculated and what does it mean?
40
Significance

• Significance of hit is measured by E-value or
  expect value
• Each alignment has a bit score (S)
• E-value is number of alignments with bit score ?
  S that you expect to find by chance
• E mn2-s
• m effective length of query
• n effective length (total of bases) of
  database

41
BLAST statistics

• bit score
• larger S, smaller E-value
• length of query
• longer queries usually generate larger E-values
• size of database
• larger database results in larger E-values

42
Calculation of raw score

• Raw score calculated from number of identities,
  mismatches, gaps and characters in the
  alignment
• R aI bX cO dG
• I number of identities a reward for each
  identity
• X number of mismatches b reward for each
  mismatch
• O number of gaps c is gap-opening penalty
• G number of d is gap-extension penalty

43
Calculation of raw score
atgctctggccacggcacttgcggatcccagggtgatctgtgcacctgcg
ata
atgctatggccacgggtcttgtggatccca---t
gatgtgtgcacctgcgata
There are 46 identities, 4 mismatches, 1 gap, and
3 - characters
R 46 (-3)(4) (5)(1) (2)(3) 23
44
Calculation of bit score

• Bit score is obtained from raw score by

?R lnK
S
ln2

• and K are normalizing parameters
• dependent on the scoring matrix

45

• ? 1.37
• K 0.711
• H 1.31
• Effective query length 34
• Effective database length 7,806,816,630

46
Calculation of bit score

• Bit score is obtained from raw score by

?R lnK
S
ln2
(1.37)(23) ln(0.711)
S
46
ln2
47
Calculation of E-value

• E mn2-s
• m effective length of query
• n effective length of the database

In this example, S 46, m 34 and n
7,806,816,630
E 0.004
48
Significance of alignment
probability that the observed match could have
happened by chance
P
number of matches as good as the observed one
that would be expected to appear by chance in a
database of the size probed Expect value
E
49
Significance of alignments

• P values between 0 and 1
• E P x size of the database
• E values range from 0 to the size of the database

50
E values

• Strongly dependent on the size of the database
• E-value from search of 9000 protein db is 100x
  smaller than E-value for exact same alignment in
  a search of a 900,000 protein db

51
Caveats

• Repetitive sequence
• Regions of low complexity
• Repeated motifs
• Unusually high number of low abundant amino acids
  (i.e. cysteines)

52
Filtering LCR
gtASH1_HUMAN Achaete-scute homolog 1
(HASH1) MESSAKMESGGAGQQPQPQPQQPFLPPAACFFATAAAAAAA
AAAAAAQSAQQQQQQQQQQQQAPQLRPAADGQPSGGGHKSAPKQVKRQRS
SSPELMRCKRRLNFSGFGYSLPQQQPAAVARRNERERNRVKLVNLGFATL
REHVPNGAANKKMSKVETLRSAVEYIRALQQLLDEHDAVSAAFQAGVLSP
TISPNYSNDLNSMAGSPVSSYSSDEGSYDPLSPEEQELLDFTNWF

• gtASH1_HUMAN Achaete-scute homolog 1 (HASH1)
  MESSAKMESGGXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
  XXXXXXXXXXXXXXXXXXXXXXQPSGGGHKSAPKQVKRQRSSSPELMRCK
  RRLNFSGFGYSLPQQQPAAXXXXXXXXXXXXXXVNLGFATLREHVPNGAA
  NKKMSKVETLRSAVEYIRALQQLLDEHDAVSAAFQAGVLSPTISPNYSND
  LNSMAGSPVSSYSSDEGSYDPLSPEEQELLDFTNWF

53
Alignment without filtering Note low E-value (E
1e-13) alignment in region 120-133.
54
Alignment with filtering turned on. Note higher
E-value (4e-7) Xs in region 120-133 as a
consequence of the filtering.
55
Flavors of BLAST

• BLASTP
• Protein query protein DB
• BLASTN
• Megablast, discontinuous megablast, blastn
• Nucleotide query, nucleotide DB
• BLASTX
• Nucleotide query translated in 3 frames protein
  DB
• TBLASTN
• Protein query, Nucleotide DB translated in 3
  frames
• TBLASTX
• Nucleotide query, nucleotide DB both translated
  in 3 frames