PROJECT TITLE: AN IMPROVED DEEP LEARNING METHOD TO
PREDICT THE ABILITY OF SGRNA OFF-TARGET PROPENSITY
Download all the required/available files from gdrive (3.5MB):
drive.google.com/drive/folders/1rDPcagh_iWWV1W9OWdIn8BeVnKdsxWSl?usp=sharing
Project Objectives:
(a)
To review prediction of deep learning method for sgRNA off-target propensity.
(b)
To model the prediction of sgRNA off-target propensity using CnnCrispr.
(c)
To measure the accuracy of the prediction model.
–
Firstly, read SGRNA PROJECT IDEA.pdf
–
Secondly, read the Reference Research Paper.pdf
–
Thirdly, run all the codes in github/google drive, If there is any errors you may fix
–
Fourthly, make sure you run from existing codes or make 5 models called CnnCrispr,
CFD, CNN_std, MIT, DeepCrispr, then run with available data or test case (Some has
only test case file) (Note: Not limited to what codes or data I have provided, you may
refer to other research papers or github links, but make sure you provide the
reference to me with brief explanation regarding why you have used this)
–
Fifthly, compare the best model among all the developed/runnable models
–
Sixthly, make difference than original code results to make differentiate with the
original coding, otherwise it will be looks like copying or stealing other’s code
Although there are original coding but we need to update coding based on the a, b, c
objectives. Make sure all the work fulfil the a, b, c objectives requirement.
Let me explain more in details:
Refer to the provided links of Gdrive/github, where you can get the dataset/test case files
and output results, or run attached coding CNN_crsipr.ipynb & cfd-score-calculator.py
where you can see whatever I have done the editing till now, and you may start from there.
For example: If require, you can get the dataset from the specific model github link datasets
for example for CNNCRISPR Model link (
github.com/LQYoLH/CnnCrispr.git
) and it’s
dataset named offtarget data.
Special Note: I have very limited time to complete all the tasks. Therefore, according to our
a, b, c objectives and the tasks are to run or develop the 5 models and calculate their
Read more here: Source link