Bioinformatics 7: Sequence alignment and its significance

Question 1

The 2 types of homolog and their differences?

Answer 1

ortholog - separated via speciation event

paralog - separated via duplication event

Question 2

What is meant by the orthology conjecture?

Answer 2

Orthologs are more likely to show more functional conservation that paralogs

i.e. ortholog genes usually related in function
paralog genes duplicate and diverge (as 2 copies of gene, same function)

Question 3

What is meant by ‘chance similarity’?

Answer 3

Any two sequences that show similarity by chance

not structurally or functionally

Question 4

2 ways dna sequences might differ?

Answer 4

Mismatches
Gaps

created by substitutions and indels

Question 5

What is a dotplot in the context of allignment?

Answer 5

Matrix of 2 sequences marked where rows and colums match

Used by alignment algorithms to find most likely evolutionary pathway between the 2 sequences

Question 6

Which is more common - indels or substitutions? How does this affect alignment?

Answer 6

Substitutions far more common than indels
-> must be considered in alignment algorithms

thus ‘quality’ of alignments is assessed via a scoring matrix (matches +ve, mismatches 0, gaps -ve) -> algorithms maximise score

Question 7

Types of gap penalty?

Answer 7

Constant
Proportional
Affine

Question 8

How would a penalty applied to an amino acid substitution vary in severity?

Answer 8

If amino acid which has been substituted is similar in chemical properties (function) -> low penalty

If completely different, likely to be deleterious -> high penalty

Question 9

How do heuristic algorithms work and why are they used over dynamic programming algorithms? Example of one?

Answer 9

Heuristic methods assume high scoring alignments contain short regions of exact matches

• > they break queries into short ‘words’ and look for matches above a threshold
• > initial hits examined to see if they can be extended
• > alignment then scored to quantify similarity
  e. g. BLAST

Question 10

How does BLAST work?

Answer 10

Basic local alignment search tool

word (W) size: 3 (proteins), 11 (DNA)
-> searches only for word matches above threshold, T

• > matches above T extended (form HSPs) until gaps cause alignment score to fall drastically
• > neighbouring HSPs are joined, HSPs in low identity regions are not joined

High-scoring segment pairs (HSPs) along query are reported + ordered by score

Question 11

Types of blast searches and their uses?

Answer 11

blastn: nucleotide query vs nucleotide db (what gene is this?)
blastp: protein query vs protein db (what protein is this?)
blastx: translated nucleotide query vs protein db (does this DNA code for a known protein?)
tblastn: protein query vs translated nucleotide db (what DNA might encode this protein)
tblastx: translated nucleotide query vs translated nucleotide db (does this DNA code for a novel protein?)

Question 12

what is FASTA? how does it compare to BLAST?

Answer 12

Heuristic algorithm
AND sequence format (single line of description followed by sequence data)

• more sensitive to distant relationships but slower than BLAST

Question 13

In the context of a search result, what is a P value and an E value?

Answer 13

P value = probability of observing as high of an alignment score between 2 unrelated sequences of the same length + composition

E (Expect) Value = How often a match would be expected to occur in a db by chance (at a given p value)