1. Mida kasutame sarnaste järjestuste leidmiseks:
BLAST
BLAST (PSI-BLAST) on optimiseeritud selleks
et leida homolooge – ühise evolutsioonilise
päritoluga järjestusi.

3. Sarnasuse mõõtmine valkudes Log-odd sarnasuste tabel
A R N D C Q E G H I L K M F P S T W Y V B Z X * A R N D C Q E G H I L K M F P S T W Y V B
Z X *

4. Sarnasuse mõõtmine nukleiinhapetes
Identity matrix: A C G T A C G T
Transition/transversion: A C G T A C G T
BLAST DNA matrix: A C G T A C G T

5. 2. Vajadused on muutunud
- Genoomi assambleerimine
- Järjestuste lokaliseerimine genoomis
- mRNA võrdlemine genoomi järjestusega
Need probleemid ei vaja nii keerukat
sarnasusmaatriksit ega afiinset gap-penalty mudelit

6. 3. SNP-de lokaliseerimine
SNPd <-> inimese genoom
50-500bp <-> bp
SNP1 chr location
SNP2 chr location
SNP3 chr location
SNP4 chr location
SNP5 chr location
SNP6 chr location
SNP7 chr location
SNP8 chr location
SNP9 chr location
SNP10 chr location
...

7. 4. Milliseid tööriistu kasutada
BLAST – väga aeglane
MEGABLAST - aeglane
SSAHA - ?
BLAT - ?
UNIMARKER (UM) - ?

8. 5. SSAHA Sequence Search and Alignment by Hashing Algorithm
Koostab tabeli (indeksi) kõigist genoomis olevatest “sõnadest”
ja jätab meelde nende asukoha genoomis
Tüüpiline sõna pikkus 10 nt.
step 1
step 2
step 10

9. Bioinformaatika rakendusi SSAHA
TTTTTTAAAAGAGAAAAAATTCTGACGGGGGCATAACTGGAGAATAAAGTGATAAAATACTGCTGAAACAAAAAGTCATCTG
Otsing:
GGGGGCATAACTGGAGAAGGAGAA
, , ,
sõnade koguarv 3*109 ,
4 bytes each = 12 GB

10. Bioinformaatika rakendusi UM
TTTTTTAAAAGAGAAAAAATTCTGACGGGGGCATAACTGGAGAATAAAGTGATAAAATACTGCTGAAACAAAAAGTCATCTG
Indeksi koostamine:
sõna pikkus 15,
kombinatsioonide arv 415 = 109 = 4 bytes
unikaalsete variantide arv =
15-mer ID location
x (4 + 4) bytes =
= 1.3 GB

14. Using this binary representation,we can process
the DNA sequence using some of the bit operations used in
computer science.For example,a left-shift operation adding 1
or 0,depending on the new nucleotide read in,will give the
next N-mer DNA (Fig.5A).Other operations to facilitate rapid
searches of,say,complementary sequences should also be
possible,although this was not explored in the present study.
The N-mers were then placed in a binary tree (Fig.5B)as
tree nodes,along with their chromosome ID, contig ID,
sequence position on the contig,and their occurrence count
and links to left child and right child,respectively,for subse-
quently encountered N-mers with the same row value but a
larger or smaller column value (Fig.5B).By traversing every
tree node after the genome scanning was completed,all UMs
of length N along with their genomic location were identified;
they are the nodes with the occurrence count equal to one.

16. Bioinformaatika rakendusi GenomeTester
TTTTTTAAAAGAGAAAAAATTCTGACGGGGGCATAACTGGAGAATAAAGTGATAAAATACTGCTGAAACAAAAAGTCATCTG
Indeksi koostamine:
sõna pikkus 16
kombinatsioonide arv 416 = 4*109
sõnade koguarv 3*109 ,
4+4 bytes each = 24 GB
16-mer ID location

17. Figure 2. GenomeTester is signficantly faster for the ‘genome test’ than any other program.
The ‘genome test’ here means finding locations of all primers (16 nt. from the 3’ end) in the human genome and calculation of possible PCR products. Tests were performed on PC-Linux based server, Pentium III, 2 GB RAM, SCSI-RAID0 hard drives. BLAST and MEGABLAST were used without dust
filter, word length was 12 for MEGABLAST and 10 for SSAHA.