4. transcriptomics

From Biolecture.org
Revision as of 05:31, 13 June 2017 by imported>Incheonkim (Created page with "<p>Date 6.2</p> <p>question : how to predict 3’UTR region and 5’ UTR region??</p> <p>-------------------------------------------</p> <p>* shape of gene</p> <p>mod...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Date 6.2

question : how to predict 3’UTR region and 5’ UTR region??

-------------------------------------------

* shape of gene

model -> 5'UTR -> exon -intron -exon - intron- exon - intron - exon -> 3'UTR

---------------------------------------------

Answer : we can predict the UTR sequence by HMM [hidden markov model]

 

concept of HMM

 

?markov chain – markov published the concepts

.if the next result was effected from something, it must be effected by previous sequence

 

let’s assume the output result is...

A-> B-> C-> D -> E.

and if it exist hidden algorithm.

there exist the possibility of the steps.

A       ->        B       ->        C

     hidden 1         hidden 2

in output : process from A to B

exist possibility A -> hidden -> B....

for A -> B -> C

possibility to have A ->B-> C order

: A x hidden 1 x B hidden2 x C...

------------------------------------------

come back to our own question...

how to predict 5’UTR & 3’ UTR?

we do have HMM & sequence information.

applying the HMM model for our own sequence

----------------------------------------

in 5',UTR region

over the species there exist

conserved pattern & difference exist

 

each position have it’s own probability in HMM model

for given sequence calculate it with position specific score..

 

making Database for the sequence
training HMM for given sequence.

---------------------------------------------------