[PDF] - Visualizing alignments DOROTHYCROWFOOTHODGKIN PDF Document

SLIDE 1

23‐Mar‐15 1

Visualizing alignments

Bas E. Dutilh Systems Biology: Bioinformatic Data Analysis Utrecht University, March 24th 2015

Dot‐plot

DOROTHYCROWFOOTHODGKIN DOROTHY--------HODGKIN

Insertions and deletions in protein structure

When comparing sequences, we sometimes observe large

insertions or deletions in otherwise similar proteins

– If the ancestral state is not known, it can be unclear if it is an insertion in one sequence, or a deletion in the other sequence – We use the word indel to We use the word indel to describe either possibility

In some cases, these indels can have a limited effect on

the structure, and thus on the function of the protein

Protein domains

Proteins often have a modular architecture consisting of

discrete structural and functional regions called domains

In many cases, different domains are encoded in different

exons

The size of domains

Average protein domain: ~100 amino acids
Most domains (90%) have <200 amino acids
Individual domains vary from 36 to 692 amino acids

Domains are like amino acid LEGO blocks

SLIDE 2

23‐Mar‐15 2

Domain re‐arrangement can yield new proteins

By using domains, evolution can make new, complex

proteins!

Discovering protein domains

Protein domains can be discovered by using bioinformatics

– Compare many protein sequences – Use local alignments – Src Homology 2

Colors allow you to visibly assess alignments

Random unaligned sequences Well‐aligned homologs

Color schemes

Color schemes are not always consistent
DNA: each nucleotide has a different color
Proteins: different colors represent different physico‐

chemical properties of the amino acids

A C G T

Motifs in protein sequences

“Motif” rhymes with “beef”
Motifs are functional units, but smaller

than protein domains

A motif is a short sequence pattern with a certain function

– Some residues in the motif can be highly conserved in evolution

An alignment of occurrences of a motif shows which

residues are more/less conserved in evolution

– The active site of Hexokinase proteins

SLIDE 3

23‐Mar‐15 3

Motifs in DNA sequences

Inside gene coding regions: conserved genetic regions
Outside coding regions: transcription factor binding sites

(TFBS)

– TATA box: found in promoters of Archaea and Eukaryotes, binds transcription factors or histones

Summarizing many aligned sequences

Sometimes it is handy to summarize

hundreds of aligned sequences

The consensus sequence is the sequence

containing the most frequent residues at each position

– The TATA box

Unknown nucleotide: N
Unknown nucleotide: N
Unknown amino acid: X

Sequence profiles

Consensus of a protein motif of 4 amino acids:
A sequence profile shows much more detail:

Position 1 Position 2 Position 3 Position 4 A C 10 D 6 E F G H 7 I 7 K L M N P 2 Q R S 4 T V 1 W Y 3

Most conserved position

Sequence profiles “summarize” biochemistry

A sequence profile represents all the possible sequences

and the sequence conservation at once

– Motifs – Protein families

This is almost like describing the real

biochemical interactions!

...atG AA AT TT C ac... ...ccG AA GT TT C tg...

We can predict that a conserved

position is important for the function

f the protein because it is rarely

changed in evolution

The profile shows:

– Which positions are more conserved – Which positions are less conserved

g ...agG AA AA TT C aa... ...gtG AA AT TT C cg... ...caG AA AT TT C tc... ...tgG AA AT TT C gt...

A 2 1 0 6 6 5 1 0 0 0 2 1 C 2 1 0 0 0 0 0 0 0 6 1 2 G 1 2 6 0 0 1 0 0 0 0 1 2 T 1 2 0 0 0 0 5 6 6 0 2 1

DNA sequence logos

At each position, possible nucleotides are shown by

stacked letters

– Letter heights relative to frequencies pi (i = A, C, G, T)

The total stack height shows the conservation

– Information content at position k (in bits):

) ( log ) 4 ( log ) (

2 , , , 2 i T G C A i i

p p k I





 

The TATA box

Maximum information in a completely conserved position (e.g. always T)

– pA = 0; pC = 0; pG = 0; pT = 1 – Assume that 0 log2 (0) = 0 I = log2 (4) + (0 + 0 + 0 + 1 log2 (1)) = 2

Minimum information in a completely unconserved position (random)

– pA = 0.25; pC = 0.25; pG = 0.25; pT = 0.25 I = 2 + (0.25 log2 (0.25) + 0.25 log2 (0.25) + 0.25 log2 (0.25) + 0.25 log2 (0.25)) I = 2 + (– 0.5 – 0.5 – 0.5 – 0.5) = 0

Transcription factor binding sites (TFBS)

Transcription factors are proteins that bind a specific DNA

sequence

SLIDE 4

23‐Mar‐15 4

Protein sequence logos

At each position, possible amino acids are shown by stacked letters

– Letter heights relative to amino acid frequencies pi (pA, pC, pD, pE, pF, pG, pH ,

pI, pK, pL, pM, pN, pP, pQ, pR, pS, pT, pV, pW, and pY)

Helix‐turn‐helix motifs

– The total stack height shows the conservation – Information content at position k (in bits):

Maximum information in a completely conserved position

I = log2 (20) + 0 = 4.3219

Minimum information in a completely random position

I = 4.3219 + (20 · (– 0.216)) = 0 ) ( log ) 20 ( log ) (

2 20 .. 1 2 i i i

p p k I





 

An exam question

) ( log ) 4 ( log ) (

2 , , , 2 i T G C A i i

p p k I





  ) ( log ) 20 ( log ) (

2 2 i i

p p k I



 

a. Which positions are fully conserved?
b. Which positions are fully random?

c. Why is the y‐axis different between the two sequence logos?

d. Give the maximum stack height for DNA sequence logos (in bits).
e. Give the maximum stack height for protein sequence logos.

f. Give both the consensus sequences. ) ( log ) 20 ( log ) (

2 20 .. 1 2 i i i

p p k I







Weblogo

Weblogo is a webserver to create sequence logos from a

multiple alignment: weblogo.berkeley.edu

Useful programs

Bioinformatic programs to align sequences:

– Clustal – T‐Coffee – MAFFT

Programs to visualize alignments:

– Clustal – Jalview – Seaview

Jalview

Sequence identifiers Aligned sequences Conservation: identity at position Consensus: frequency of top residue Quality: conservation of similar amino acids

Weighing conservation of a position in an alignment

Sequence alignments that use

information about the sequence conservation at each position into

...atG AA AT TT C ac... ...ccG AA GT TT C tg... ...agG AA AA TT C aa... tG AA AT TT C

account are called profile alignments

In profile alignments, the important

(conserved) residues have a bigger impact on the alignment score

– More conserved residues are weighed higher in the similarity score – Less conserved residues are weighed lower in the similarity score

→ Profile alignments are more sensive than sequence alignments

...gtG AA AT TT C cg... ...caG AA AT TT C tc... ...tgG AA AT TT C gt...

A 2 1 0 6 6 5 1 0 0 0 2 1 C 2 1 0 0 0 0 0 0 0 6 1 2 G 1 2 6 0 0 1 0 0 0 0 1 2 T 1 2 0 0 0 0 5 6 6 0 2 1