CSE700 Parallel Programming Assignment 6

1
CSE-700 Parallel ProgrammingAssignment 6
???

• POSTECH
• Oct 19, 2007

2
Species and Sequences
Species
Sequence 1
Sequence 2
...
Sequence n
3
Ortholog
Last Common Ancestor
S
By speciation
Human
Dog
S1
S2
4
Paralog
Human
S
By duplication
Human
S1
S1'
5
Inparalog
Last Common Ancestor
S
By speciation
Human
Chimpanzee
S1
S2
S1'
By duplication
6
Paralog - Outparalog
LCA Last Common Ancestor
LCA
Human
Dog
7
Coortholog
Species A
Species B
S1'
S1
S2
S2'
8
Input

• Assume a total of n species S1, S2, ..., Sn
• For each pair of species Si, Sj
• Ortholog and paralog relations
• Thus n(n 1)/2 ortholog/paralog files

9
Seed Ortholog
Species A
Species B
Cluster
1.0
Si
Sj
10
Invariant No Two Seed Orthologs for Any Sequence
Species A
Species B
Sj
1.0
Si
1.0
Sk
11
Ortholog and Paralogs
Species A
Species B
Cluster
1.0
Si
Sj
Si'
12
Output

• Assume a total of n species S1, S2, ..., Sn
• Ortholog and paralog relations among all these
  species
• In each cluster,
• seed ortholog from each pair of species
• paralogs may be included.

13
Example of Cluster 1
A
B
S2
S2'
C
D
S3
S3'
14
Example of Cluster 2
A
B
S2
S2'
C
D
S3
S3'
15
Bad Clusters 1
A
B
S2
S2'
C
D
S3
S3'
E
16
Bad Clusters 2
C
D
S3
S4'
S4
S4''
S5
E
17
Input File Format

• Each line consists of
• Cluster number
• Similarity score
• Species name
• Seed ortholog
• Sequence name

18
Goal

• Implement ANY sequential algorithm
• There is no definitive answer.
• Then parallelize it.
• A parser and an output module are provided.
• no string comparion
• all integer operations