Multiple sequence alignment

Transcription

1 Multiple sequence alignment Wednesday, October 11, 2006 Sarah Wheelan

2 Copyright notice Many of the images in this powerpoint presentation are from Bioinformatics and Functional Genomics by J Pevsner (ISBN ). Copyright 2003 by Wiley. These images and materials may not be used without permission from the publisher. Visit

3 Multiple sequence alignment: today s goals to define what a multiple sequence alignment is and how it is generated; to describe profile HMMs to introduce databases of multiple sequence alignments to introduce ways you can make your own multiple sequence alignments to show how a multiple sequence alignment provides the basis for phylogenetic trees Page 319

4 Multiple sequence alignment: definition Wallace, Blackshields, Higgins Curr Opp Struct Biol :

5 Multiple sequence alignment: properties Generally align proteins: nucleotides less well-conserved nucleotide sequences are less informative -> harder to align with high confidence How do you know if you have the correct alignment of a protein family? Is there one correct alignment? for two proteins sharing 30% amino acid identity, about 50% of the individual amino acids are superposable in the two structures Page 320

6 Proportion of structurally superposable residues in pairwise alignments as a function of sequence identity Proportion of residues in common core Globin Cytochrome c Serine protease Immunoglobulin domain Sequence identity (%) After Chothia & Lesk (1986)

7 Multiple sequence alignment: features some aligned residues, such as cysteines that form disulfide bridges, may be highly conserved there may be conserved motifs such as a transmembrane domain there may be conserved secondary structure features there may be regions with consistent patterns of insertions or deletions (indels), often indicating the presence of subfamilies Page 320

8 Multiple sequence alignment: uses MSA is more sensitive than pairwise alignment in detecting homologs Detect conserved motifs and domains (from database search or from alignment of a family) Phylogenetic analysis: MSA provides accurate estimation of evolutionary distances 2 and 3 structure prediction: MSA predictions significantly better than from a single sequence very few bioinformatics protocols bypass MSA stage Page 321

9 Multiple sequence alignment: methods There are four main ways to make a multiple sequence alignment: (1) Exact methods (2) Progressive alignment (CLUSTALW, MUSCLE) (3) Iterative approaches (Praline, IterAlign) (4) Consistency-based methods (ProbCons)

10 Multiple sequence alignment: methods Exact methods: dynamic programming Instead of the 2-D dp matrix that you saw in the Needleman-Wunsch technique, think about a 3-D, 4-D etc matrix. Exact methods give optimal alignments but are not feasible in time or space for more than ~10 sequences Still an extremely active field though the holy grail

11 Multiple sequence alignment: methods Progressive methods: use a guide tree (a little like a phylogenetic tree but NOT a phylogenetic tree) to determine how to combine pairwise alignments one by one to create a multiple alignment. Making multiple alignments using trees was a very popular subject in the 80s. Fitch and Yasunobu (1974) may have first proposed the idea, but Hogeweg and Hesper (1984) and many others worked on the topic before Feng and Doolittle (1987) hit the scene they made one important contribution that got their names attached to this alignment method. Examples: CLUSTALW, MUSCLE

12 Multiple sequence alignment: methods Iterative methods: compute a sub-optimal solution and keep modifying that intelligently using dynamic programming or other methods until the solution converges. Examples: IterAlign, Praline, MAFFT

13 Multiple sequence alignment: methods Consistency-based algorithms: generally use a database of both local high-scoring alignments and long-range global alignments to create a final alignment These are very powerful, very fast, and very accurate methods Examples: T-COFFEE, Prrp, DiAlign, ProbCons

14 Multiple sequence alignment: methods Aaack! This is confusing! How do we know which program to use? There are benchmarking multiple alignment datasets that have been aligned painstakingly by hand, by structural similarity, or by extremely time- and memory-intensive automated exact algorithms. Also, to be practical, some programs have interfaces that are much more user-friendly than others. And most programs are excellent so it depends on your preference. If your proteins have 3D structures, USE THESE to help you judge your alignments.

15 Multiple sequence alignment: methods Benchmarking tests suggest that ProbCons, a consistency-based/progressive algorithm, performs the best on the BAliBASE set, although MUSCLE, a progressive alignment package, is an extremely fast and accurate program. CLUSTALW, everyone s old favorite, continues to be a standout and is included in almost every MSA paper you will see. It has withstood the test of time. Plus, it has a nice interface (expecially with CLUSTALX) and is easy to use.

16 Multiple sequence alignment: methods Example of progressive MSA using CLUSTALW: two data sets Five distantly related lipocalins (human to E. coli) Five closely related RBPs When you do this, obtain the sequences of interest in FASTA format! Page 321

17 Get sequences from GenBank

18 Get sequences from GenBank

19 Use Clustal W to do a progressive MSA

20 Feng-Doolittle MSA occurs in 3 stages [1] Do a set of global pairwise alignments (Needleman and Wunsch) [2] Create a guide tree [3] Progressively align the sequences according to the guide tree Page 321

21 Progressive MSA stage 1 of 3: generate global pairwise alignments five distantly related lipocalins best score

22 Progressive MSA stage 1 of 3: generate global pairwise alignments five closely related lipocalins best score

23 Number of pairwise alignments needed Need to align every sequence to every other sequence (handshake problem) For N sequences, (N-1)(N)/2 For 5 sequences, (4)(5)/2 = 10 Page 322

24 Feng-Doolittle stage 2: guide tree Convert similarity scores to distance scores (see text) A guide tree shows the distance between objects Use UPGMA (defined in Chapter 11) ClustalW provides a syntax to describe the tree A guide tree is not a phylogenetic tree Page 323

25 Progressive MSA stage 2 of 3: generate a guide tree calculated from the distance matrix five distantly related lipocalins

26 Progressive MSA stage 2 of 3: generate guide tree five closely related lipocalins

27 Feng-Doolittle stage 3: progressive alignment Make a MSA based on the order in the guide tree Start with the two most closely related sequences Then add the next closest sequence Continue until all sequences are added to the MSA Rule: once a gap, always a gap. <- this is what made them famous! Page 324

28 Why once a gap, always a gap? There are many possible ways to make a MSA Where gaps are added is a critical question Gaps are first added to the first two (closest) sequences To change the initial gap choices later on would be to give more weight to distantly related sequences To maintain the initial gap choices is to trust that those gaps are most believable (shouldn t let the final alignment be affected most by distantly related sequences!) Page 324

29 Progressive MSA stage 3 of 3: progressively align the sequences following the branch order of the tree Fig Page 324

30 Progressive MSA stage 3 of 3: CLUSTALX output

31 Clustal W alignment of 5 closely related lipocalins CLUSTAL W (1.82) multiple sequence alignment gi pir A39486 MEWVWALVLLAALGSAQAERDCRVSSFRVKENFDKARFSGTWYAMAKKDP 50 gi sp P18902 RETB_BOVIN ERDCRVSSFRVKENFDKARFAGTWYAMAKKDP 32 gi ref NP_ MKWVWALLLLAAW--AAAERDCRVSSFRVKENFDKARFSGTWYAMAKKDP 48 gi sp Q00724 RETB_MOUS MEWVWALVLLAALGGGSAERDCRVSSFRVKENFDKARFSGLWYAIAKKDP 50 gi sp P04916 RETB_RAT MEWVWALVLLAALGGGSAERDCRVSSFRVKENFDKARFSGLWYAIAKKDP 50 ********************:* ***:***** gi pir A39486 EGLFLQDNIVAEFSVDENGHMSATAKGRVRLLNNWDVCADMVGTFTDTED 100 gi sp P18902 RETB_BOVIN EGLFLQDNIVAEFSVDENGHMSATAKGRVRLLNNWDVCADMVGTFTDTED 82 gi ref NP_ EGLFLQDNIVAEFSVDETGQMSATAKGRVRLLNNWDVCADMVGTFTDTED 98 gi sp Q00724 RETB_MOUS EGLFLQDNIIAEFSVDEKGHMSATAKGRVRLLSNWEVCADMVGTFTDTED 100 gi sp P04916 RETB_RAT EGLFLQDNIIAEFSVDEKGHMSATAKGRVRLLSNWEVCADMVGTFTDTED 100 *********:*******.*:************.**:************** gi pir A39486 PAKFKMKYWGVASFLQKGNDDHWIIDTDYDTYAAQYSCRLQNLDGTCADS 150 gi sp P18902 RETB_BOVIN PAKFKMKYWGVASFLQKGNDDHWIIDTDYETFAVQYSCRLLNLDGTCADS 132 gi ref NP_ PAKFKMKYWGVASFLQKGNDDHWIVDTDYDTYAVQYSCRLLNLDGTCADS 148 gi sp Q00724 RETB_MOUS PAKFKMKYWGVASFLQRGNDDHWIIDTDYDTFALQYSCRLQNLDGTCADS 150 gi sp P04916 RETB_RAT PAKFKMKYWGVASFLQRGNDDHWIIDTDYDTFALQYSCRLQNLDGTCADS 150 ****************:*******:****:*:* ****** ********* Fig Page 326

32 Progressive MSA stage 3 of 3: progressively align the sequences following the branch order of the tree: Order matters THE LAST FAT CAT THE FAST CAT THE VERY FAST CAT THE FAT CAT THE LAST FAT CAT THE FAST CAT --- THE LAST FA-T CAT THE FAST CA-T --- THE VERY FAST CAT Adapted from C. Notredame, Pharmacogenomics 2002 THE LAST FA-T CAT THE FAST CA-T --- THE VERY FAST CAT THE ---- FA-T CAT

33 Progressive MSA stage 3 of 3: progressively align the sequences following the branch order of the tree: Order matters THE FAT CAT THE FAST CAT THE VERY FAST CAT THE LAST FAT CAT THE FA-T CAT THE FAST CAT THE ---- FA-T CAT THE ---- FAST CAT THE VERY FAST CAT Adapted from C. Notredame, Pharmacogenomics 2002 THE ---- FA-T CAT THE ---- FAST CAT THE VERY FAST CAT THE LAST FA-T CAT

34 Progressive MSA stage 3 of 3: progressively align the sequences following the branch order of the tree: CLUSTALW results

35 MUSCLE: next-generation progressive MSA 1) Build a draft progressive alignment Determine pairwise similarity through k-mer counting (not by alignment) Compute distance (triangular distance) matrix Construct tree using UPGMA Construct draft progressive alignment following tree

36 MUSCLE: next-generation progressive MSA 2) Improve the progressive alignment Compute pairwise identity through current MSA Construct new tree with Kimura distance measures Compare new and old trees: if improved, repeat this step, if not improved, then we re done

37 MUSCLE: next-generation progressive MSA 3) Refinement of the MSA Split tree in half by deleting one edge Make profiles of each half of the tree Re-align the profiles Accept/reject the new alignment

38

39 MUSCLE output (formatted with SeaView)

40 Praline output for the same alignment: pure iterative approach

41 ProbCons consistency-based approach Combines iterative and progressive approaches with a unique probabilistic model. Uses Hidden Markov Models (more in a minute) to calculate probability matrices for matching residues, uses this to construct a guide tree Progressive alignment hierarchically along guide tree Post-processing and iterative refinement (a little like MUSCLE)

42 ProbCons consistency-based approach

43 ProbCons output for the same alignment: how consistency iteration helps

44 Multiple sequence alignment to profile HMMs in 90 s people began to see that aligning sequences to profiles gave much more information than pairwise alignment alone. Hidden Markov models (HMMs) are states that describe the probability of having a particular amino acid residue at arranged In a column of a multiple sequence alignment HMMs are probabilistic models Like a hammer is more refined than a blast, an HMM gives more sensitive alignments than traditional techniques such as progressive alignments Page 325

45 Simple Markov Model Rain = dog may not want to go outside Sun = dog will probably go outside 0.85 S Markov condition = no dependency on anything but nearest previous state ( memoryless ) 0.2 R

46 Simple Hidden Markov Model 0.15 S P(dog goes out in rain) = 0.1 P(dog goes out in sun) = 0.85 R Observation: YNNNYYNNNYN 0.2 (Y=goes out, N=doesn t go out) What is underlying reality (the hidden state chain)?

47 An HMM is constructed from a MSA Example: five lipocalins GTWYA (hs RBP) GLWYA (mus RBP) GRWYE (apod) GTWYE (E Coli) GEWFS (MUP4) Fig Page 327

48 GTWYA GLWYA GRWYE GTWYE GEWFS Prob p(g) 1.0 p(t) 0.4 p(l) 0.2 p(r) 0.2 p(e) p(w) 1.0 p(y) 0.8 p(f) 0.2 p(a) 0.4 p(s) 0.2 Fig Page 327

49 GTWYA GLWYA GRWYE GTWYE GEWFS Prob p(g) 1.0 p(t) 0.4 p(l) 0.2 p(r) 0.2 p(e) p(w) 1.0 p(y) 0.8 p(f) 0.2 p(a) 0.4 p(s) 0.2 P(GEWYE) = (1.0)(0.2)(1.0)(0.8)(0.4) = log odds score = ln(1.0) + ln(0.2) + ln(1.0) + ln(0.8) + ln(0.4) = Fig Page 327

50 GTWYA GLWYA GRWYE GTWYE GEWFS P(GEWYE) = (1.0)(0.2)(1.0)(0.8)(0.4) = log odds score = ln(1.0) + ln(0.2) + ln(1.0) + ln(0.8) + ln(0.4) = G:1.0 T:0.4 L:0.2 R:0.2 E:0.2 W:1.0 Y:0.8 F:0.2 E:0.4 A:0.4 S:0.2 Fig Page 327

51 Structure of a hidden Markov model (HMM) p 4 p 2 I x p 1 M p 3 p 5 p 6 I y p 7

52 Structure of a hidden Markov model (HMM): Trellis representation Fig Page 328

53 HMMER: build a hidden Markov model Determining effective sequence number... done. [4] Weighting sequences heuristically... done. Constructing model architecture... done. Converting counts to probabilities... done. Setting model name, etc.... done. [x] Constructed a profile HMM (length 230) Average score: bits Minimum score: bits Maximum score: bits Std. deviation: bits Fig Page 329

54 HMMER: calibrate a hidden Markov model HMM file: lipocalins.hmm Length distribution mean: 325 Length distribution s.d.: 200 Number of samples: 5000 random seed: histogram(s) saved to: [not saved] POSIX threads: HMM : x mu : lambda : max : Fig Page 329

55 HMMER: search an HMM against GenBank Scores for complete sequences (score includes all domains): Sequence Description Score E-value N gi ref XP_ (XM_129259) ret e gi sp P04916 RETB_RAT Plasma retinol e gi ref XP_ (XM_005907) sim e gi ref NP_ (NM_006744) ret e gi sp P02753 RETB_HUMAN Plasma retinol e gi ref NP_ (NC_003197) out e-90 1 gi ref NP_ : domain 1 of 1, from 1 to 195: score 454.6, E = 1.7e-131 *->mkwvmkllllaalagvfgaaerdafsvgkcrvpspprgfrvkenfdv mkwv++llllaa + +aaerd Crv+s frvkenfd+ gi MKWVWALLLLAA--W--AAAERD------CRVSS----FRVKENFDK 33 erylgtwyeiakkdprferglllqdkitaeysleehgsmsataegrirvl +r++gtwy++akkdp E GL+lqd+I+Ae+S++E+G+Msata+Gr+r+L gi ARFSGTWYAMAKKDP--E-GLFLQDNIVAEFSVDETGQMSATAKGRVRLL 80 enkelcadkvgtvtqiegeasevfltadpaklklkyagvasflqpgfddy +N+++cAD+vGT+t++E dpak+k+ky+gvasflq+g+dd+ gi NNWDVCADMVGTFTDTE DPAKFKMKYWGVASFLQKGNDDH 120 Fig Page 330

56 HMMER: search an HMM against GenBank match to a bacterial lipocalin gi ref NP_ : domain 1 of 1, from 1 to 177: score 318.2, E = 1.9e-90 *->mkwvmkllllaalagvfgaaerdafsvgkcrvpspprgfrvkenfdv M+LL+ +A a ++ Af+v++C++p+PP+G++V++NFD+ gi MRLLPVVA------AVTA-AFLVVACSSPTPPKGVTVVNNFDA 36 erylgtwyeiakkdprferglllqdkitaeysleehgsmsataegrirvl +rylgtwyeia+ D+rFErGL + +ta+ysl++ +G+i+V+ gi KRYLGTWYEIARLDHRFERGL---EQVTATYSLRD DGGINVI 75 enkelcadkvgtvtqiegeasevfltadpaklklkyagvasflqpgfddy Nk++++D EG+a ++t+ P +++lk+ Sf++p++++y gi NKGYNPDR-EMWQKTEGKA---YFTGSPNRAALKV----SFFGPFYGGY 116 Fig Page 330

57 HMMER: search an HMM against GenBank Scores for complete sequences (score includes all domains): Sequence Description Score E-value N gi sp P27485 RETB_PIG Plasma retinol e gi pir A39486 plasma retinol e gi ref XP_ (XM_129259) ret e gi sp P04916 RETB_RAT Plasma retinol e gi ref XP_ (XM_005907) sim e gi sp P02753 RETB_HUMAN Plasma retinol e gi ref NP_ (NM_006744) ret e gi ref NP_ : domain 1 of 1, from 1 to 199: score 600.2, E = 2.6e-175 *->mewvwalvllaalggasaerdcrvssfrvkenfdkarfsgtwyaiak m+wvwal+llaa+ a+aerdcrvssfrvkenfdkarfsgtwya+ak gi MKWVWALLLLAAW--AAAERDCRVSSFRVKENFDKARFSGTWYAMAK 45 KDPEGLFLqDnivAEFsvDEkGhmsAtAKGRvRLLnnWdvCADmvGtFtD KDPEGLFLqDnivAEFsvDE+G+msAtAKGRvRLLnnWdvCADmvGtFtD gi KDPEGLFLQDNIVAEFSVDETGQMSATAKGRVRLLNNWDVCADMVGTFTD 95 tedpakfkmkywgvasflqkgnddhwiidtdydtfavqyscrllnldgtc tedpakfkmkywgvasflqkgnddhwi+dtdydt+avqyscrllnldgtc gi TEDPAKFKMKYWGVASFLQKGNDDHWIVDTDYDTYAVQYSCRLLNLDGTC 145 Fig Page 330

58 Databases of multiple sequence alignments BLOCKS CDD DOMO INTERPRO iproclass MetaFAM PFAM PRINTS PRODOM PROSITE SMART Page

59 Databases of multiple sequence alignments BLOCKS (HMM) CDD (HMM) DOMO (Gapped MSA) INTERPRO iproclass MetaFAM Pfam (profile HMM library) PRINTS PRODOM (PSI-BLAST) PROSITE SMART Page

60 Databases of multiple sequence alignments BLOCKS CDD DOMO INTERPRO iproclass MetaFAM PFAM PRINTS PRODOM PROSITE SMART Integrative resources Page

61 Pfam (protein family) database: or

62 Pfam (protein family) database

63 Pfam (protein family) database

64 Pfam (protein family) database

65 Pfam (protein family) database

66 SMART: Simple Modular Architecture Research Tool (emphasis on cell signaling) Search for conserved domains Page 338

67

68

69 Databases of multiple sequence alignments BLOCKS CDD DOMO INTERPRO iproclass MetaFAM PFAM PRINTS PRODOM PROSITE SMART CDD at NCBI = PFAM + SMART Page 333

70 CDD uses RPS-BLAST: reverse position-specific Query = your favorite protein Database = set of many PSSMs CDD is related to PSI-BLAST, but distinct CDD searches against profiles generated from pre-selected alignments Purpose: to find conserved domains in the query sequence You can access CDD via DART at NCBI Page 333

71 CDART (from NCBI) allows CDD searches

72

73 BLOCKS server BLOCKS: ungapped MSA Very highly conserved regions

74

75

76

77

78

79

80 Fig Page 341

81

82 See chapter 8

83 PROSITE has a syntax to define signatures See chapter 8

84 PopSet: polymorphisms in population data Fig Page 342

85 PopSet: polymorphisms in population data Fig Page 342

86 PopSet: polymorphisms in population data

87 Big Picture sites very useful eview-edital-desc.html Nice overview site lots of alignment utilities, including editors and viewers. Most are current. Pedro s molecular biology tools. Most are current.

88 MSA databases: manual vs. automated curation Manual curation: Pfam PROSITE BLOCKS PRINTS Automated curation: DOMO PRODOM MetaFam + comprehensive - alignment errors

89 Multiple sequence alignment programs AMAS CINEMA ClustalW ClustalX DIALIGN HMMT Match-Box MultAlin MSA Musca PileUp SAGA T-COFFEE

90 Selecting sequences for a PileUp in GCG Fig Page 346

91 GCG PileUp Fig Page 346

92 GCG PileUp Fig Page 347

93 Multiple sequence alignment algorithms Local Global Progressive PIMA CLUSTAL PileUp other Iterative DIALIGN SAGA Fig Page 348

94 Strategy for assessment of alternative multiple sequence alignment algorithms [1] Create or obtain a database of protein sequences for which the 3D structure is known. Thus we can define true homologs using structural criteria. [2] Try making multiple sequence alignments with many different sets of proteins (very related, very distant, few gaps, many gaps, insertions, outliers). [3] Compare the answers. Page 346

95 Strategy for assessment of alternative multiple sequence alignment algorithms Commonly used benchmarking databases: BAliBASE SMART SABmark PREFAB (Protein Reference Alignment Benchmark)

96 Conclusions: assessment of alternative multiple sequence alignment algorithms [1] As percent identity among proteins drops, performance (accuracy) declines also. This is especially severe for proteins < 25% identity. Proteins <25% identity: 65% of residues align well Proteins <40% identity: 80% of residues align well Page 350

97 Conclusions: assessment of alternative multiple sequence alignment algorithms [2] Orphan sequences are highly divergent members of a family. Surprisingly, orphans do not disrupt alignments. Page 350

98 Conclusions: assessment of alternative multiple sequence alignment algorithms [3] Separate multiple sequence alignments can be combined (e.g. RBPs and lactoglobulins). Profile-based methods are very powerful but the strength of the profile depends on the quality of the initial alignment. Page 350

99 Conclusions: assessment of alternative multiple sequence alignment algorithms [4] When proteins have large N-terminal or C-terminal extensions, local alignment algorithms are superior. PileUp (global) is an exception. Page 350