Parallel Smith-Waterman Algorithm - Local Sequence Alignment - - - PowerPoint PPT Presentation

▶

Mar 10, 2024 343 likes •500 views

Parallel Smith-Waterman Algorithm - Local Sequence Alignment - myson @ postech.ac.kr CSE700-PL @ POSTECH Programming Language Laboratory p.1/14 Outline Smith-Waterman algorithm How to parallelize? Parallel Smith-Waterman algorithm A

SLIDE 1

Parallel Smith-Waterman Algorithm

Local Sequence Alignment -

myson @ postech.ac.kr CSE700-PL @ POSTECH

Programming Language Laboratory – p.1/14

SLIDE 2

Outline

Smith-Waterman algorithm How to parallelize? Parallel Smith-Waterman algorithm A wavefront algorithm A scalable algorithm Variants of a wavefront algorithm

Programming Language Laboratory – p.2/14

SLIDE 3

Smith-Waterman Algorithm

Local sequence alignment Sequence similarity comparison ex) S=ATAGCT and T=GATATGCA S =

ATA-GCT

T = GATATGCA

Programming Language Laboratory – p.3/14

SLIDE 4

Smith-Waterman Algorithm (cont.)

1. Making a similarity matrix
2. Identifying the optimal alignment

− − G A T A T G C A − A −1 −1 −1 T −2 −2 A −1 −3 −1 −1 −1 G −1 −1 −2 −2 C −1 −3 −1 T −1 −1 −1 −2

Programming Language Laboratory – p.4/14

SLIDE 5

Similarity Matrix

|S| = n, |T| = m

Programming Language Laboratory – p.5/14

SLIDE 6

Similarity Matrix (cont.)

#define Match (-1) #define Mismatch 1 #define Gap 2 Score( i, j ) = ( S[ i ] == T[ j ] ? Match : Mismatch ) A[ i ][ j ] = min (0, A[ i-1 ][ j ] + Gap, A[ i ][ j-1 ] + Gap, A[ i-1 ][ j-1 ] + Score( i, j ))

Programming Language Laboratory – p.6/14

SLIDE 7

Similarity Matrix (cont.)

A[ i ][ j ] = min (0, A[ i-1 ][ j ] + Gap, A[ i ][ j-1 ] + Gap, A[ i-1 ][ j-1 ] + Score( i, j ))

Programming Language Laboratory – p.7/14

SLIDE 8

How to parallelize?

Any idea. . . ?

Programming Language Laboratory – p.8/14

SLIDE 9

Parallel SW - Algorithms

A wavefront algorithm

Programming Language Laboratory – p.9/14

SLIDE 10

Parallel SW - Algorithms (cont.)

A scalable algorithm

Programming Language Laboratory – p.10/14

SLIDE 11

Parallel SW - Algorithms (cont.)

Variants of a wavefront algorithm - 1

Programming Language Laboratory – p.11/14

SLIDE 12

Parallel SW - Algorithms (cont.)

Variants of a wavefront algorithm - 2

Programming Language Laboratory – p.12/14

SLIDE 13

How to parallelize?

Any idea. . . ?

Programming Language Laboratory – p.13/14

SLIDE 14

End

Any Questions. . . ?

Programming Language Laboratory – p.14/14