Multiple Sequence Alignment

Alignment can be easy or difficult

GCGGCCCA GCGGCCCA GCGTTCCA GCGTCCCA GCGGCGCA ******** TTGACATG TTGACATG TTGACATG TTGACATG TTGACATC ******** TCAGGTAGTT TCAGGTAGTT TCAGCTGGTT TCAGCTAGTT TTAGCTAGTT ********** CCGGGG---A CCGGTG--GT -CTAGG---A -CTAGGGAAC -CTCTG---A ?????????? GGTGG GGTGG GGTGG GGTGG GGTGA *****
Easy

Difficult due AACCG AAGCC to insertions ACGCG or deletions ACGCG (indels) ACGCG *****

Homology: Definition
Homology: similarity that is the result of inheritance from a common ancestor - identification and analysis of homologies is central to phylogenetic systematics. An Alignment is an hypothesis of positional homology between bases/Amino Acids.

Multiple Sequence AlignmentGoals

To generate a concise, information-rich summary of sequence data. Sometimes used to illustrate the dissimilarity between a group of sequences. Alignments can be treated as models that can be used to test hypotheses. Does this model of events accurately reflect known biological evidence.

Alignment of 16S rRNA can be guided by secondary structure

<---------------(--------------------HELIX 19---------------------) <---------------(22222222-000000-111111-00000-111111-0000-22222222 Thermus ruber UCCGAUGC-UAAAGA-CCGAAG=CUCAA=CUUCGG=GGGU=GCGUUGGA Th. thermophilus UCCCAUGU-GAAAGA-CCACGG=CUCAA=CCGUGG=GGGA=GCGUGGGA E.coli UCAGAUGU-GAAAUC-CCCGGG=CUCAA=CCUGGG=AACU=GCAUCUGA Ancyst.nidulans UCUGUUGU-CAAAGC-GUGGGG=CUCAA=CCUCAU=ACAG=GCAAUGGA B.subtilis UCUGAUGU-GAAAGC-CCCCGG=CUCAA=CCGGGG=AGGG=UCAUUGGA Chl.aurantiacus UCGGCGCU-GAAAGC-GCCCCG=CUUAA=CGGGGC=GAGG=CGCGCCGA match ** *** * ** ** * **

Alignment of 16S rRNA sequences from different bacteria

Protein Alignment may be guided by Tertiary Structure Interactions

Escherichia coli DjlA protein

Homo sapiens DjlA protein

Multiple Sequence AlignmentMethods

3 main methods of alignment:

Manual Automatic Combined

Manual Alignment - reasons

Might be carried out because: Alignment is easy. There is some extraneous information (structural). Automated alignment methods have encountered the local minimum problem. An automated alignment method can be improved.

Dynamic programming
2 methods: Dynamic programming
Consider 2 protein sequences of 100 amino acids in length. If it takes 1002 seconds to exhaustively align these sequences, then it will take 1003 seconds to align 3 sequences, 1004 to align 4 sequences...etc. More time than the universe has existed to align 20 sequences exhaustively.

Progressive alignment

Progressive Alignment
Devised by Feng and Doolittle in 1987. Essentially a heuristic method and as such is not guaranteed to find the optimal alignment. Requires n-1+n-2+n-3...n-n+1 pairwise alignments as a starting point Most successful implementation is Clustal (Des Higgins)

Overview of ClustalW Procedure

Hbb_Human Hbb_Horse Hba_Human Hba_Horse Myg_Whale 1 2 3 4 5 .17 .59 .59 .77

CLUSTAL W
.60 .59 .77 .13 .75 .75

Quick pairwise alignment: calculate distance matrix

4

Hbb_Human Hbb_Horse Hba_Human Hba_Horse Myg_Whale

1 2

Neighbor-joining tree (guide tree)

alpha-helices
1 2 3 4 5 PEEKSAVTALWGKVN--VDEVGG GEEKAAVLALWDKVN--EEEVGG PADKTNVKAAWGKVGAHAGEYGA AADKTNVKAAWSKVGGHAGEYGA EHEWQLVLHVWAKVEADVAGHGQ 1 2 3 4

Progressive alignment following guide tree

ClustalW- Pairwise Alignments

First perform all possible pairwise alignments between each pair of sequences. There are (n-1)+(n-2)...(nn+1) possibilities. Calculate the distance between each pair of sequences based on these isolated pairwise alignments. Generate a distance matrix.

Path Graph for aligning two sequences.

Possible alignment
1 1 0 1 0 -1 Score for this path= 2 Scoring Scheme: Match: +1 Mismatch: 0 Indel: -1

Alignment using this path

1 1 0 1 0 -1

GATTCGAATTC

Optimal Alignment 1
1 1 -1 1 1 1 Alignment using this path GA-TTC GAATTC

Alignment score: 4

Optimal Alignment 2
1 -1 1 1 1 1 Alignment using this path G-ATTC GAATTC

Alignment score: 4

ClustalW- Guide Tree

Generate a Neighbor-Joining guide tree from these pairwise distances. This guide tree gives the order in which the progressive alignment will be carried out.

Neighbor joining method

The neighbor joining method is a greedy heuristic which joins at each step, the two closest sub-trees that are not already joined. It is based on the minimum evolution principle. One of the important concepts in the NJ method is neighbors, which are defined as two taxa that are connected by a single node in an unrooted tree Node 1

What is required for the Neighbour joining method?

Distance matrix

Distance Matrix
Rice 84.9 0.0 117.8 122.4 122.6 Mosquito 105.6 117.8 0.0 84.7 80.8 Monkey 90.8 122.4 84.7 0.0 3.3 Human 86.3 122.6 80.8 3.3 0.0

PAM Spinach Rice Mosquito Monkey Human

Spinach 0.0 84.9 105.6 90.8 86.3

First Step
PAM distance 3.3 (Human - Monkey) is the minimum. So we'll join Human and Monkey to MonHum and we'll calculate the new distances.

Mon-Hum

Mosquito

Spinach

Rice Human

Monkey

Calculation of New Distances

After we have joined two species in a subtree we have to compute the distances from every other node to the new subtree. We do this with a simple average of distances:
Dist[Spinach, MonHum] = (Dist[Spinach, Monkey] + Dist[Spinach, Human])/2 = (90.8 + 86.3)/2 = 88.55

Mon-Hum

Spinach

Human

Monkey

Next Cycle
PAM Spinach Rice Mosquito MonHum Spinach 0.0 84.9 105.6 88.6 Rice 84.9 0.0 117.8 122.5 Mosquito 105.6 117.8 0.0 82.8 MonHum 88.6 122.5 82.8 0.0

Mos-(Mon-Hum) Mon-Hum Rice Spinach Mosquito Human Monkey

Penultimate Cycle
PAM Spinach Rice MosMonHum Spinach 0.0 84.9 97.1 Rice 84.9 0.0 120.2 MosMonHum 97.1 120.2 0.0

Mos-(Mon-Hum) Spin-Rice Mon-Hum Mosquito Human Monkey

Rice

Spinach

Last Joining
PAM Spinach MosMonHum SpinRice 0.0 108.7 MosMonHum 108.7 0.0

(Spin-Rice)-(Mos-(Mon-Hum))

Mos-(Mon-Hum) Spin-Rice Mon-Hum Spinach Mosquito Human Monkey

Rice

Unrooted Neighbor-Joining Tree

Human Spinach Monkey Mosquito

Rice

Multiple Alignment- First pair

Align the two most closely-related sequences first. This alignment is then fixed and will never change. If a gap is to be introduced subsequently, then it will be introduced in the same place in both sequences, but their relative alignment remains unchanged.

ClustalW- Decision time

Next, consult the guide tree to see what alignment is performed next.
Align a third sequence to the first two Or Align two entirely different sequences to each other.

Option 1

Option 2

ClustalW- Alternative 1
If the situation arises where a third sequence is aligned to the first two, then when a gap has to be introduced to improve the alignment, each of these two entities are treated as two single sequences.

ClustalW- Alternative 2
If, on the other hand, two separate sequences have to be aligned together, then the first pairwise alignment is placed to one side and the pairwise alignment of the other two is carried out.

ClustalW- Progression
The alignment is progressively built up in this way, with each step being treated as a pairwise alignment, sometimes with each member of a pair having more than one sequence.

 Advantages: Speed.

ClustalW-Good points/Bad points

Disadvantages: No objective function. No way of quantifying whether or not the alignment is good No way of knowing if the alignment is correct.

ClustalW-Local Minimum
Potential problems: Local minimum problem. If an error is introduced early in the alignment process, it is impossible to correct this later in the procedure. Arbitrary alignment.

Increasing the sophistiaction of the alignment process.

Should we treat all the sequences in the same way? - even though some sequences are closely-related and some sequences are distant relatives. Should we treat all positions in the sequences as though they were the same? - even though they might have different functions and different locations in the 3-dimensional structure.

ClustalW- Caveats
Sequence weighting Varying substitution matrices Residue-specific gap penalties and reduced penalties in hydrophilic regions (external regions of protein sequences), encourage gaps in loops rather than in core regions. Positions in early alignments where gaps have been opened receive locally reduced gap penalties to encourage openings in subsequent alignments

Sequence weighting
First we must be able to categorise sequences according to whether they have close relatives or if they are distantly-related to the other sequences (calculated directly from the guide tree). Weights are normalised, so that the largest weight is 1. Closely-related sequences have a large amount of the same information, so they are downweighted. These weights are multiplication factors.

ClustalW- User-supplied values

Two penalties are set by the user (there are default values, but you should know that it is possible to change these). GOP- Gap Opening Penalty is the cost of opening a gap in an alignment. GEP- Gap Extension Penalty is the cost of extending this gap.

ClustalW- Manipulation of penalties

Although GOP and GEP are set by the user, the program attempts to manipulate these according to the following criteria:
Dependence on the weight matrix: Dependence on the similarity of the sequences: The percent identity of the sequences is used as a scaling factor to increase the GOP for closely-related sequences and decrease it for more distantly-related sequences.

ClustalW
Dependence on the length of the sequences:
The program uses the formula
GOP->(GOP+log(MIN(N,M))*(Average residue mismatch score)*(percent identity scaling factor)

The logarithm of the length of the shortest sequence is used as a scaling factor to increase the GOP with increasing length

Dependence on the difference in lengths of the two sequences:

GEP-> GEP*(1.0+|log(N/M)|)

Position-Specific gap penalties

Before any pair of (groups of) sequences are aligned, a table of GOPs are generated for each position in the two (sets of) sequences. The GOP is manipulated in a position-specific manner, so that it can vary over the sequences. If there is a gap at a position, the GOP and GEP penalties are lowered, the other rules do not apply. This makes gaps more likely at positions where gaps already exist.

Discouraging too many gaps

If there is no gap opened, then the GOP is increased if the position is within 8 residues of an existing gap. This discourages gaps that are too close together. At any position within a run of hydrophilic residues, the GOP is decreased. These runs usually indicate loop regions in protein structures. A run of 5 hydrophilic residues is considered to be a hydrophilic stretch. The default hydrophilic residues are: D, E, G, K, N, Q, P, R, S But this can be changed by the user.

Divergent Sequences
The most divergent sequences (most different, on average from all of the other sequences) are usually the most difficult to align. It is sometimes better to delay their aligment until later (when the easier sequences have already been aligned). The user has the choice of setting a cutoff (default is 40% identity). This will delay the alignment until the others have been aligned.

Advice on progressive alignment

Progressive alignment is a mathematical process that is completely independent of biological reality. Can be a very good estimate Can be an impossibly poor estimate. Requires user input and skill. Treat cautiously Can be improved by eye (usually) Often helps to have colour-coding. Depending on the use, the user should be able to make a judgement on those regions that are reliable or not. For phylogeny reconstruction, only use those positions whose hypothesis of positional homology is unimpeachable

Alignment of protein-coding DNA sequences

It is not very sensible to align the DNA sequences of protein-coding genes.

ATGCTGTTAGGG ATGCTCGTAGGG

ATGCT-GTTAGGG ATGCTCGTA-GGG

The result might be highly-implausible and might not reflect what is known about biological processes. It is much more sensible to translate the sequences to their corresponding amino acid sequences, align these protein sequences and then put the gaps in the DNA sequences according to where they are found in the amino acid alignment.

Manual Alignment- software

GDE- The Genetic Data Environment (UNIX) CINEMA- Java applet available from:
http://www.biochem.ucl.ac.uk

Seqapp/Seqpup- Mac/PC/UNIX available from:

http://iubio.bio.indiana.edu

SeAl for Macintosh, available from:

http://evolve.zoo.ox.ac.uk/Se-Al/Se-Al.html

BioEdit for PC, available from:

http://www.mbio.ncsu.edu/RNaseP/info/programs/BIOEDIT/bi oedit.html