Bioinformatics

What is BLOSUM62 and why is it useful?

BLOSUM62 is one of the most common substituation matrixes used to score amino acid matches in sequence alignments. It is useful because all aino acid changes (mutations) are not equally probable

What is a Genbank?

A database of all known nucleotide sequences

What is a Uniprot?

A database of all known protein sequences

What does it mean when two sequences are homologs?

That they have a common ancestor

What is an E-value?

E-values are used to assess the significance of sequencce database searches. An E-value for a given hit is the number of sequences that by chance would get a better score, Typically you like the E-value to be <0,001 to be significant

What is BLAST and what is it best for?

• Heuristic tool for finding local sequence similarity

• Fast, but may miss weak/distant matches

• Works by findning short exact matches (“words”) and extending them

What is PSI-BLAST and how does it differ from BLAST?

• Iterative version of BLAST for proteins

• Builds a position-specific scoring matrix (PSSM)

• More sensitive than BLAST; detects distant homologs

• Slower than BLAST duo to iterations

What is Smith-Waterman and why is it important?

• Exact method for local sequence alignment using dynamic programming

• Guarantees the optimal alignment

• Very sensitive, but slowest

Rank BLAST, PSI-BLAST and Smith-Waterman by speed

BLAST > PSI-BLAST > Smith-Waterman

Rank BLAST, PSI-BLAST and Smith-Waterman by sensitivity

Smith-Waterman > PSI-BLAST > BLAST

What is PSSM?

• Position Specific Scoring Matrix

• Represents conservation of residues at each position in a sequence alignment

• Scores indicate how likely a residue is at a postion compared to background

How is a PSSM constructed?

1. Start with a query sequence

2. BLAST search to find similar sequences

3. Align hits to the query

4. Count residue frequencies at each position

5. Convert frequencies to log-odds scores (observed vs backround)

6. Optionally adjust for small sample sizes with pseudo-counts

What is PSSM used for?

• Detect distant homologs

• Iterative searches in PSI-BLAST

• Identify motifs/domains in proteins

• Score new sequences against known patterns