- Cal State LA - Instructional Web Server

Download Report

Transcript - Cal State LA - Instructional Web Server

ProInt Finder to Search
Protein Interactions
Shwe S. Lin
Mentor: Matteo Pellegrini, UCLA
Outlines
 Project overview
 Purpose
 Background
 My part
 Algorithms
 Tool used
 Examples
 Acknowledgments
 References
Project Overview
To develop a protein-protein interaction
database and an interactive web-based
interface to search for protein
interactions from the database
Purpose of the Project
 To study protein-protein interactions
•
Example:
p53-MDM2 interactions
 To predict protein functions which may
be inferred from analyzing protein
interactions
•
Example:
protein A interacts with 5 cell cycle proteins
and we therefore infer that it is a cell cycle
protein
Importance of Studying Protein
Interactions
Example:
 Essential for cell communications which result in
activation or inactivation of biological responses
• Protein MDM2 inactivates p53’s function as a
tumor suppressor
• EPO interacts with the EPO receptor to trigger
growth of erythrocytes
Traditional Methods for Studying
Protein Functions
• Sequence alignment techniques
• Protein’s 3-D structure
Limitations of Traditional Methods
 Yield functional information only on
experimentally characterized homologous
proteins
 Detect protein’s biochemical function only;
not biological process
Approach to Address Limitations
 Develop protein interaction databases
• Example: DIP, MINT, BIND etc.
 Implement methods for function prediction
• Example: guilt-by-association
My Part
 Develop web application (ProInt Finder)
to utilize the database for studying
protein interactions
• Django: Python Web framework
• Database: Automated collections of data by
text mining of web
Algorithm (Input)
 Accept query: gene name or gene id
of interest
• Example - Gene name: p53
Gene ID: 7157
Algorithm (Search)
actors
acted by
T
a
Q
IF
re
uF
ic
ne
k
(U
T
en
im
dc
eo
e
dm
™
to
pa
re
sn
e
s
d
e
sa
e
th
dis
)d
pe
ic
co
tum
rep
. ressor
 Search for gene_ids from the database gene table
Algorithm (Output)
 Return:
gene_1_id interacts with gene_2_id
 Result:
List of protein pairs that are experimentally identified
to interact with each other
pro_ A >> activates pro_B
>> acetylates pro_C
>> methylates pro_D
>> interacts with pro_E
pro_E phosphorylates >> pro_A
pro_H binds to
pro_I activates
pro_K regulates
>>
>>
>>
Tool Used: Django
• Python web framework to build Web applications
• Provides Model-View-Controller (MVC)
approach to programming
• Database layer: Models or data models
• Controller layer: View or control logic
• View layer: Templates or user interface
Database Layer: Models
• Each model is a Python class
• Contains fields and behaviors of the stored data
• Each model maps to a single database table
A Model in Django
from django.db import models
class Gene(models.Model):
gene_id = models.IntegerField (primary_key=True)
long_name = models.TextField( )
symbol = models.TextField( )
tax_id = models.ForeignKey( Species )
class Meta:
db_table = 'protInt_gene'
ordering = ['symbol']
Gene
Gene ID
Gene Name
Taxonomy ID
Fields/ class attributes
View Layer: Templates
•
A text file that contains variables
and tags
ProInt Finder
• Variables: Get replaced with
values
when the template is evaluated
• Tags: Control the logic of the template
Templates for each of the web pages for ProInt Finder
index.html
search.html
results.html
Control Layer
• Takes user input from search.html template
• Defines how to process the data
• Returns results to results.html template
User input
search.html
Controller
(View)
results.html
Protein interactions
Home Page
This is a link which leads to the search page.
These are links to other protein interaction databases.
Search page
Result page
User Input
Description of the protein
Query protein interacts with different protein
Proteins interacts with query protein
More on Result
Web Links
Click on the link
Acknowledgements




Dr. Matteo Pellegrini
Shawn Cokus
Joseph Kim and Cory Tobin
Dr. Sandra Sharp and Dr. Wendie
Johnston
 SoCalBSI
 NIH, NSF, and LAOC
References
• http://www.djangoproject.com/
Questions
?