Topic 3: PAM, BLOSUM, and Scoring Matrices

0.0(0)
Studied by 0 people
call kaiCall Kai
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
GameKnowt Play
Card Sorting

1/35

encourage image

There's no tags or description

Looks like no tags are added yet.

36 Terms

1
New cards
Why are biologically meaningful scoring matrices needed?
Amino acid substitutions have different biological likelihoods, so they should not all receive the same score.
2
New cards
What is a scoring matrix?
A table assigning scores to aligned nucleotide or amino acid pairs.
3
New cards
What is a substitution?
Replacement of one nucleotide or amino acid by another.
4
New cards
What is a diagonal matrix element?
A score for aligning a residue with itself, such as A with A.
5
New cards
What is an off-diagonal element?
A score for aligning two different residues.
6
New cards
What does a symmetric scoring matrix mean?
S(i,j)=S(j,i).
7
New cards
What are the two models behind log-odds scoring?
Related/common-ancestor model and unrelated/random model.
8
New cards
What is the related-sequence model?
Aligned residue pairs are assumed to derive from a common ancestor.
9
New cards
What is the random model?
Aligned residues are assumed unrelated and occur by background frequencies.
10
New cards
What is the random-model probability of ordered residues i and j?
Q(i,j)=p(i)×p(j).
11
New cards
What is the related-model probability given in class?
P(i,j)=p(j)×P(i|j).
12
New cards
What is a log-odds substitution score?
The logarithm of the ratio of related-model probability to random-model probability.
13
New cards
What does a positive substitution score mean?
The pair is more common under the related model than expected by chance.
14
New cards
What does a zero substitution score mean?
The pair occurs about as often as expected by chance.
15
New cards
What does a negative substitution score mean?
The pair is less common under the related model than expected by chance.
16
New cards
What does PAM stand for?
Point Accepted Mutation.
17
New cards
Who developed PAM matrices?
Margaret Dayhoff and colleagues.
18
New cards
How are PAM matrices derived?
Use closely related proteins, infer accepted mutations, estimate a PAM1 substitution model, then extrapolate using matrix powers.
19
New cards
What does PAM1 represent?
One accepted point mutation per 100 amino acids on average.
20
New cards
What does XX mean in PAM-XX?
Evolutionary distance in PAM1 steps.
21
New cards
What does a larger PAM number mean?
Greater evolutionary distance; generally useful for more divergent sequences.
22
New cards
What does PAM250 mean?
The PAM1 mutation probability matrix is applied for 250 evolutionary steps, not 250 observed differences per 100 sites.
23
New cards
What does BLOSUM stand for?
BLOcks SUbstitution Matrix.
24
New cards
Who developed BLOSUM matrices?
Henikoff and Henikoff.
25
New cards
How are BLOSUM matrices derived?
Count amino acid pairs in conserved ungapped blocks, cluster highly similar sequences, calculate observed and expected pair frequencies, then log-odds scores.
26
New cards
Does BLOSUM use an explicit evolutionary tree?
No; it uses observed substitutions in conserved protein blocks.
27
New cards
What does XX mean in BLOSUM-XX?
The sequence identity threshold used to cluster similar sequences.
28
New cards
What does BLOSUM62 mean?
Sequences above 62% identity are grouped when building the matrix.
29
New cards
What does a larger BLOSUM number generally mean?
A matrix more appropriate for closely related sequences.
30
New cards
Which matrices are generally used for more divergent sequences?
Higher PAM numbers and lower BLOSUM numbers.
31
New cards
What is observed pair frequency q(i,j)?
Count of amino acid pair i,j divided by the total pair count.
32
New cards
What is expected frequency for an identical unordered pair?
e(i,i)=p(i)^2.
33
New cards
What is expected frequency for different unordered pairs?
e(i,j)=2p(i)p(j).
34
New cards
What is the BLOSUM half-bit score equation?
S(i,j)=round[2 log2(q(i,j)/e(i,j))].
35
New cards
How do you calculate sequence identity?
(Number of identical aligned residue pairs / number of compared aligned positions)×100%.
36
New cards