1. Sequence homology and alignment
Lessons 3-4
Sequence homology and alignment

2. Homology
Similarity between characters due to a common ancestry

3. Sequence homology
Similarity between sequences that results from a common ancestor
VLSPAVKWAKVGAHAAGHG
VLSEAVLWAKVEADVAGHG
Basic assumption:
Sequence homology → similar structure/function

4. Sequence alignment
Alignment: Comparing two (pairwise) or more (multiple) sequences. Searching for a series of identical or similar characters in the sequences.

5. Homology Ortholog – homolog with similar function (via speciation)
Paralog – homolog which arose from gene duplication
Common use:
Orthologs –
2 homologs from different species
Paralogs – 2 homologs within the same species

6. How close? Rule of thumb:
Proteins are homologous if 25% identical (length >100)
DNA sequences are homologous if 70% identical

7. Twilight zone
< 20% identity in proteins – may be homologous and may not be….
(Note that 5% identity will be obtained completely by chance!)

8. Why sequence alignment?
Predict characteristics of a protein –
use the structure/function of known proteins for predicting the structure/function of an unknown protein

9. Sequence modifications
Sequences change in the course of evolution due to random mutations
Three types of mutations:
Insertion - an insertion of a letter or several letters to the sequence. AAGA AAGTA
Deletion - deleting a letter (or more) from the sequence.
AAGA AGA
Substitution - replacing a sequence letter by another.
AAGA AACA
Insertion or Deletion = Indel

10. Local vs. Global
Global alignment: forces alignment in regions which differ
Global alignment – finds the best alignment across the entire two sequences.
Local alignment – finds regions of similarity in parts of the sequences.
ADLGAVFALCDRYFQ
|||| |||| |
ADLGRTQN-CDRYYQ
Local alignment will return only regions of good alignment
ADLG CDRYFQ
|||| |||| |
ADLG CDRYYQ

11. When global and when local?

12. Global alignment
PTK2 protein tyrosine kinase 2 of human and rhesus monkey

13. Protein tyrosine kinase domain

14. Protein tyrosine kinase domain
Human PTK2 and leukocyte tyrosine kinase
Both function as tyrosine kinases, in completely different contexts
Ancient duplication

15. Global alignment of PTK and LTKX

16. Local alignment of PTK and LTK

17. Pairwise alignment AAGCTGAATTCGAA AGGCTCATTTCTGA
One possible alignment:
AAGCTGAATT-C-GAA
AGGCT-CATTTCTGA-

18. AAGCTGAATT-C-GAA
AGGCT-CATTTCTGA-
This alignment includes:
2 mismatches
4 indels (gap)
10 perfect matches

19. Choosing an alignment:
Many different alignments are possible:
AAGCTGAATTCGAA
AGGCTCATTTCTGA
A-AGCTGAATTC--GAA
AG-GCTCA-TTTCTGA-
AAGCTGAATT-C-GAA
AGGCT-CATTTCTGA-
Which alignment is better?

20. Alignment scoring - scoring of sequence similarity:
Assumes independence
Each position considered separately
Score at each position
Positive if identical
Negative if different or gap
Score of an alignment is the sum of position scores
Can be positive or negative

21. Example - naïve scoring system:
Perfect match: +1
Mismatch: -2
Indel (gap): -1
AAGCTGAATT-C-GAA
AGGCT-CATTTCTGA-
A-AGCTGAATTC--GAA
AG-GCTCA-TTTCTGA-
Score: = (+1)x10 + (-2)x2 + (-1)x4 = 2
Score: = (+1)x9 + (-2)x2 + (-1)x6 = -1
Higher score  Better alignment

22. Scoring system: The choice of +1,-2, and -1 scores is quite arbitrary
Different scoring systems  different alignments
Scoring systems implicitly represent a particular theory of evolution
Some mismatches are more plausible
Transition vs. Transversion
LysArg ≠ LysCys
Gap extension Vs. Gap opening

23. Scoring matrix
Representing the scoring system as a table or matrix n  n (n is the number of letters the alphabet contains. n=4 for nucleotides, n=20 for amino acids)
symmetric T C G A 2 -6

24. DNA scoring matrices Uniform substitution in all nucleotides: T C G A
From To 2 -6
Match
Mismatch

25. DNA scoring matrices
Can take into account biological phenomena such as:
Transition-transversion

26. Amino Acid Scoring Matrices
Take into account physico-chemical properties

27. Amino Acid Substitutions Matrices
Actual substitutions:
Based on empirical data
Commonly used by many bioinformatic programs
PAM & BLOSUM

28. Protein matrices – actual substitutions
The idea: Given an alignment of a large number of closely related sequences we can score the relation between amino acids based on how frequently they substitute each other
M G Y D E
M G Y E E
M G Y Q E
In the fourth column
E and D are found in 7 / 8

29. PAM Matrix - Point Accepted Mutations
Based on a database of 1,572 changes in 71 groups of closely related proteins (85% identity)
Alignment was easy
Counted the number of the substitutions per amino acid pair (20 x 20)
Found that common substitutions occurred between chemically similar amino acids

30. PAM Matrices Family of matrices PAM 80, PAM 120, PAM 250
The number on the PAM matrix represents evolutionary distance
Larger numbers are for larger distances

31. Example: PAM 250
Similar amino acids have greater score

32. PAM - limitations Only one original dataset
Examining proteins with few differences (85% identity)
Based mainly on small globular proteins so the matrix is biased

33. BLOSUM
Henikoff and Henikoff (1992) derived a set of matrices based on a much larger dataset
BLOSUM observes significantly more replacements than PAM, even for infrequent pairs

34. BLOSUM: Blocks Substitution Matrix
Based on BLOCKS database
~2000 blocks from 500 families of related proteins
Families of proteins with identical function
Blocks are short conserved patterns of aa long without gaps
AABCDA----BBCDA
DABCDA----BBCBB
BBBCDA-AA-BCCAA
AAACDA-A--CBCDB
CCBADA---DBBDCC
AAACAA----BBCCC

35. BLOSUM
Each block represent sequences alignment with different identity percentage
For each block the amino-acid substitution rates were calculated to create BLOSUM matrix

36. BLOSUM Matrices
BLOSUMn is based on sequences that shared at least n percent identity
BLOSUM62 represents closer sequences than BLOSUM45

37. Example : Blosum62 derived from block where the sequences
clustered share at least 62% identity

38. More distant sequences
PAM vs. BLOSUM
PAM100 = BLOSUM90
PAM120 = BLOSUM80
PAM160 = BLOSUM60
PAM200 = BLOSUM52
PAM250 = BLOSUM45
More distant sequences

39. Scoring system =
substitution matrix +
gap penalty

40. Gap penalty We penalize gaps Scoring for gap opening & for extension:
Gap-extension penalty < gap-open penalty

41. Optimal alignment algorithms
Needleman-Wunsch (global)
Smith-Waterman (local).

42. Alignment Search Space
The “search space” (number of possible gapped alignments) for optimally aligning two sequences is exponential in the length of the sequences, n.
If n=100, there are
= = 100,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000 different alignments.
Average protein length is about n=250!

43. Searching databases

44. Searching a database
Using a sequence as the query to find homologous sequences in the database

45. DNA or protein?
For coding sequences, we can use the DNA sequence or the protein sequence to search for similar sequences.
Which is preferable?

46. Protein is better!
Selection (and hence conservation) works on the protein level: CTTTCA = Leu-Ser TTGAGT = Leu-Ser

47. Query type Nucleotides: a four letter alphabet
Amino acids: a twenty letter alphabet
Two random DNA sequences will share on average 25% of identity
Two random protein sequences will share on average 5% of identity

48. Conclusions
Using the amino acid sequence is preferable for homology search
Why use a nucleotide sequence after all?
No ORF found, e.g. newly sequenced genome
No similar protein sequences were found
Specific DNA databases are available (EST)

49. Some terminology
Query sequence - the sequence with which we are searching
Hit – a sequence found in the database, suspected as homologous

50. How do we search a database?
Assume we perform pairwise alignment of the query against all the sequences in the database
Exact pairwise alignment is O(mn) ≈ O(n2) (m – length of sequence 1, n – length of sequence 2)

51. How much time will it take?
O(n2) computations per search.
Assume n=200, so we have 40,000 computations per search
Size of database - ~60 million entries
2.4 x 1012 computations for each sequence search we perform!
Assume each computation takes 10-6 seconds  24,000 seconds ≈ 6.66 hours for each sequence search
150,000 searches (at least!!) are performed per day

52. Conclusion
Using the exact comparison pairwise alignment algorithm between query and all DB entries – too slow

53. Heuristic
Definition: a heuristic is a design to solve a problem that does not provide exact solution (but is not too bad) and reduces the time complexity of the exact solution

54. BLAST BLAST - Basic Local Alignment and Search Tool
A heuristic for searching a database for similar sequences

55. DNA or Protein All types of searches are possible. Query: DNA Protein
Database: DNA Protein
blastn – nuc vs. nuc blastp – prot vs. prot blastx – translated query vs. protein database tblastn – protein vs. translated nuc. DB tblastx – translated query vs. translated database
Translated databases:
trEMBL genPept

56. BLAST - underlying hypothesis
The underlying hypothesis: when two sequences are similar there are short ungapped regions of high similarity between the two
The heuristic:
Discard irrelevant sequences
Perform exact local alignment with remaining sequences

57. How do we discard irrelevant sequences quickly?
Divide the database into words of length w (w = 3 for protein and w = 7 for DNA)
Save the words in a look-up table that can be searched quickly
WTD
TDF
DFG
FGY
GYP …
WTDFGYPAILKGGTAC

58. BLAST: discarding sequences
When the user gives a query sequence, divide it also into words
Search the database for consecutive neighbor words

59. Neighbour words
neighbor words are defined according to a scoring matrix (e.g. BLOSUM62 for proteins) with a certain cutoff level
GFC (20)
GFB
GPC (11)
WAC (5)

60. Search for consecutive words
Neighbor word
Look for a seed: hits on the same diagonal which can be connected A
At least 2 hits on the same diagonal with distance which is smaller than a predetermined cutoff
Database record
This is the filtering stage – many unrelated hits are filtered, saving lots of time!
Query

61. Perform local alignment
Database record
Query

62. Try to extend alignment
Stop extending when the score of the alignment drops X beneath maximal score obtained this far
Discard segments with score < S
ASKIOPLLWLAASFLHNEQAPALSDAN
JWQEOPLWPLAASOIHLFACNSIFYAS
Score=15
Score=17
Score=14
X=4

63. The result – local alignment
The result of BLAST will be a series of local alignments between the query and the different hits found

64. Theoretically, we could trust any result with an E-value ≤ 1
The number of times we will theoretically find an alignment with a score ≥ Y of a random sequence vs. a random database
Theoretically, we could trust any result with an E-value ≤ 1
In practice – BLAST uses estimations. E-values of 10-4 and lower indicate a significant homology. E-values between 10-4 and 10-2 should be checked (similar domains, maybe non-homologous). E-values between 10-2 and 1 are suspicious…

65. Filtering low complexity
Low complexity regions : e.g., Proline rich areas (in protein), Alu repeats (in DNA)
Regions of low complexity generate high score of alignment BUT – this does not indicate homology

66. Solution
In BLAST there is masking of low-complexity regions in the query sequence (such regions are represented as XXXXX in query)

67. Multiple sequence alignment

68. Like pairwise alignment BUT compare n sequences instead of 2
VTISCTGSSSNIGAG-NHVKWYQQLPG
VTISCTGTSSNIGS--ITVNWYQQLPG
LRLSCSSSGFIFSS--YAMYWVRQAPG
LSLTCTVSGTSFDD--YYSTWVRQPPG
PEVTCVVVDVSHEDPQVKFNWYVDG--
ATLVCLISDFYPGA--VTVAWKADS--
AALGCLVKDYFPEP--VTVSWNSG---
VSLTCLVKGFYPSD--IAVEWWSNG--
Like pairwise alignment BUT compare n sequences instead of 2
Rows represent individual sequences
Columns represent ‘same’ position
May be gaps in some sequences

69. MSA & Evolution
MSA can give you a picture of the forces that shape evolution!
Important amino acids or nucleotides are not “allowed” to mutate
Less important positions change more easily

70. Conserved positions
Columns where all the sequences contain the same amino acids or nucleotides
Important for the function or structure
VTISCTGSSSNIGAG-NHVKWYQQLPG
VTISCTGSSSNIGS--ITVNWYQQLPG
LRLSCTGSGFIFSS--YAMYWYQQAPG
LSLTCTGSGTSFDD-QYYSTWYQQPPG

71. Consensus Sequence
The consensus sequence holds the most frequent character of the alignment at each column T G C A T G C A

72. PSSM – Position Specific Score Matrix
Profile 6 5 4 3 2 1 .
0.67 A
0.33 T C G T G C A
Profile =
PSSM – Position Specific Score Matrix

73. Alignment methods Progressive alignment (Clustal)
Iterative alignment (mafft, muscle)
All methods today are an approximation strategy (heuristic algorithm), yield a possible alignment, but not necessarily the best one

74. Progressive alignmentB C D
First step:
Compute the pairwise alignments for all against all (6 pairwise alignments)
the similarities are stored in a table D C B A 11 1 3 10 2

75. Second step: D C B A 11 1 3 10 2
cluster the sequences to create a tree (guide tree):
Represents the order in which pairs of sequences are to be aligned
similar sequences are neighbors in the tree
distant sequences are distant from each other in the tree
The guide tree is imprecise and is NOT the tree which truly describes the relationship between the sequences! A D C B

76. Third step: A B C D Align most similar pairs
Align the alignments as if each of them was a single sequence (replace with a single consensus sequence or use a profile)

77. Alignment of alignments
M Q T F
L H T W
L Q S W X
M Q T - F
L H T - W
L Q S - W
L - T I F
M - T I W
L T I F
M T I W Y

78. Iterative alignment A B C D Pairwise distance table Guide tree MSA
Iterate until the MSA doesn’t change D C B A 11 1 3 10 2
Guide tree
MSA A D C B

79. Searching for remote homologs
Sometimes BLAST isn’t enough.
Large protein family, and BLAST only gives close members. We want more distant members
PSI-BLAST
Profile HMMs

80. Construct profile from blast results
PSI-BLAST
Position Specific Iterated BLAST
Regular blast
Construct profile from blast results
Blast profile search
Final results

81. PSI-BLAST
Advantage: PSI-BLAST looks for seq.s that are close to ours, and learns from them to extend the circle of friends
Disadvantage: if we found a WRONG sequence, we will get to unrelated sequences (contamination). This gets worse and worse each iteration

82. Profile HMM Similar to PSI-BLAST: also uses a profile
Takes into account:
Dependence among sites (if site n is conserved, it is likely that site n+1 is conserved  part of a domain
The probability of a certain column in an alignment

83. PSI BLAST vs profile HMM
Less exact
Faster
More exact
Slower

84. Case study: Using homology searching
The human kinome

85. Kinases and phosphatases

86. Multi-tasking enzymes
Signal transduction
Metabolism
Transcription
Cell-cycle
Differentiation
Function of nervous and immune system …
And more

87. How many kinases in the human genome?
1970’s, advent of cloning and sequencing
Predicted: 1000 kinases in the human genome

88. How many kinases in the human genome?
2001 – human genome sequence …
As well – databases of Genbank, Swissprot, and dbEST
How do we know how many kinases are out there?

89. The human kinome
In 2002, Manning, Whyte, Martinez, Hunter and Sudarsanam set out to:
Search and cross-reference all these databases for all kinases
Characterize all found kinases

90. ePKs and aPKs Eukaryotic protein kinases (majority)
Atypical protein kinases
Sequence homology of the catalytic domain; additional regulatory domains are non-homologous
No sequence homology to ePKs; some aPK subfamilies have structural similarity to ePKs

91. The search
Several profiles were built: based on the catalytic domain of (a) 70 known ePKs (b) each subfamily of known aPKs
HMM-profile searches and PSI-BLAST searches were performed

92. The result…
478 apKs
40 ePKs
Total of 518 kinases in the human genome (half of the prediction in the 1970’s)

93. Classifying the kinases
Classification based on the catalytic domain
Classification based on the regulatory domains
189 sub-families of kinases

94. Comparison to other species
209 subfamilies of ePKs in human, worm, yeast and fly

95. The human genome has x2 kinases (in number) as fly or worm
The human genome has x2 kinases (in number) as fly or worm. Many are aPKs.
Most of them are receptor tyrosine kinases (RTKs)
Nervous system
Immune system
Hemopoiesis

96. The discovery of new kinases: a new front for battling human diseases

97. Correlating with human diseases
160 kinases mapped to amplicons seen in tumors
80 kinases mapped to amplicons in other major illnesses
Usually kinases are over-expressed in cancer and other diseases

98. Correlating with human diseases
6 kinase inhibitors have been approved till today for the use against cancer
>70 other inhibitors are in clinical trial