Sequence Searching

Hybridisation capture

• probes designed and attached to array/bead

• DNA sample added.

• Complementary sequences bind

• wash

• elute DNA from probes and sequence

Sequence searching

Computational hybridisation capture experiment

Pairwise sequence alignment

In silico comparison of two sequences using a scoring scheme (+1 for match, -1 for mismatch, -1 for gap)

BLOSUM62

common scoring scheme for aligning protein sequences

BLAST

Basic Local Alignment Search Tool, a fast sequence similarity search tool

problems with BLAST

doesn’t necessarily produce optimal alignments. Results vary based on BLAST program used, database being searched and parameters

BLAST algorithm

Divided into seeding, extension and evaluation

neighbourhood

words of the same length whose aligned score is greater than or equal to some score threshold

seeding

BLAST assumes significant matches have words in common and finds word hits in database index.

Seeding: the two-hit algorithm

BLAST looks for diagonals of word hit matches. More word hits on diagonal = more significant match.

Word hit

A match between a query word and a database index word, plus its 'neighbourhood'

Extension

Extending high-scoring seeds in each direction, adding/subtracting from the original seed score until drop-off from peak greater than threshold.

Evaluation

Calculating alignment quality using the score S

What is k in the equation E = kmne^(-λS)?

minor constant

What is m in the equation E = kmne^(-λS)?

query size

What is n in the equation E = kmne^(-λS)?

database size

What is λ in the equation E = kmne^(-λS)?

scaling factor

What is S in the equation E = kmne^(-λS)?

alignment score

What is E in the equation E = kmne^(-λS)?

number of alignments with same or larger S expected in database of same size and letter frequency, if sequences were random.

What happens to the E value when there is a small change in the S value?

large change in E value