Bioinformatics Test 2

Front 1

Smith-Waterman

Back 1

modded N&W for local. top row & left column are 0s. Any negative score is a 0.

Front 2

FASTA

Back 2

heuristic - word search (by word length "ktup") then does smith waterman. init1, init 2, opt.

Front 3

BLAST

Back 3

word search (word size W and word hit score T). The program tries to extend out matches, and LCRs are ignored.

Front 4

S value

Back 4

the alignment score

Front 5

P value

Back 5

the probability that a sequence could achieve "S" by chance

Front 6

E value

Back 6

number of sequences that could achieve "S" by chance.

Front 7

PSSM

Back 7

position specific scoring matrix - substitution scores for each AA possibility within the expected motif. (hidden markov models)

Front 8

PSI-BLAST

Back 8

uses a PSSM after the search round. This makes BLAST better than FASTA

Front 9

Feng-Doolittle

Back 9

is a MSA tool. does pair alignment --->score--->map distance--->tree

Front 10

Clustal

Back 10

is a MSA tool. aligns the most closely related sequences first, and builds a map of relationships.

Front 11

ClustalW

Back 11

clustal aligns the most closely related sequences first, then builds a map of the relationships. "W" or weightage means that more distantly related sequences count less, and more closely related sequences have higher gap penalties.

Front 12

Phi Φ

Back 12

angle of rotation between N and C

Front 13

Psi Ψ

Back 13

angle of rotation between C(alpha) and C(carboxy)

Front 14

alpha helix

Back 14

Φ = Ψ = -60

HH hh

Front 15

beta sheet

Back 15

Φ = -135
Ψ = +135

EE ee

Front 16

Chou-Fasman

Back 16

protein structure predictor.
p(a)
p(b)
p(turn)
overlapping regions

Front 17

GOR method

Back 17

protein structure predictor. Better than chou-fasman, because it incorporates "neighborhood evaluation". Uses this in combination with PSSM.