Skip to main content

Deep Biology Program



About Deep Biology Program

Cere Labs is happy to start the Deep Biology program under the umbrella of CoE with Patkar-Varde College, Goregaon. This unique program brings together multiple departments in Patkar-Varde College, Goregaon to collaborate with CereLabs. The objective is to use Deep Learning and Machine Learning for Drug Discovery and Personalised Oncology.

The Deep Biology program took place in four phases:

Phase I - April ‘17 to May ‘17 - Decide Areas  

In the first phase the following two areas were decided:
Drug Discovery and Personalised Oncology
Drug design is an expensive process. A new drug takes 10 to 15 years and costs more than $250 billion to introduce it to market. Applying Machine Learning to drug discovery will reduce both the time and cost of discovering a new drug.

Personalized oncology 
Personalized oncology is the method of offering customized medicine for a cancer patient based on the person’s genetic makeup. Machine Learning techniques accelerates the process of finding accurate treatment.

Phase II - May ‘17 to June ‘17 - Training & Assignments
Students from Bioinformatics and Computer Science went through a seven days workshop on Bioinformatics and Machine Learning. This workshop helped them to start their research in drug discovery and personalized oncology.

Phase III - June ‘17 to September ‘17 - Literature survey and decide project topic

Following two projects were finalized

Project 1:Design chemical entity suitable for inhibition for HIV-1 Protease by combination machine learning techniques & structure based drug designing.

Description: Understanding the pathway of HIV virus and identifying important drug target (i.e. HIV-1 Protease) & validating active site in protein. Approved drug parameters are retrieved from DrugBank or PubChem. Creating analogs or similar structure and checking its activity using insilico tools. Combining data of approved and similar structure suitable for applying supervised machine learning technique and generate model/equation. Retrieving the parent molecule from collected data and performing lead optimization derive a new molecule. New molecule can be tested through the equation generated by machine learning to check activity/inactivity of molecule on HIV-1 Protease.

Expected Outcome: Determine parameters for best suited for chemical entity on selected protein target & model the structure of chemical entity for further analysis.

Project 2: Identifying Drug Candidate for multidrug resistance tuberculosis using drug repositioning method & machine learning.

Description: Machine learning is used to find patterns from gene expressions retrieved from GEO database which helps in identifying differential gene expression in healthy and diseased sample. Drugs are linked with gene expression to find enrichment score for each drug. Score above 30% indicates optimal drug suitable for further optimization and testing.

Expected Outcome: Identifying drug candidate from previously drugs, optimize the drug to reduce timeline of treatment. 

Phase IV -  September ‘17 onwards - Actual Working on project
 
Students have started on the projects. The task is of collecting data and training it using Machine Learning algorithms.
 

Comments

Popular posts from this blog

Understanding Generative Adversarial Networks - Part II

In "Understanding Generative Adversarial Networks - Part I" you gained a conceptual understanding of how GAN works. In this post let us get a mathematical understanding of GANs.
The loss functions can be designed most easily using the idea of zero-sum games. 
The sum of the costs of all players is 0. This is the Minimax algorithm for GANs
Let’s break it down.
Some terminology: V(D, G) : The value function for a minimax game E(X) : Expectation of a random variable X, also equal to its average value D(x) : The discriminator output for an input x from real data, represents probability G(z): The generator's output when its given z from the noise distribution D(G(z)): Combining the above, this represents the output of the discriminator when 
given a generated image G(z) as input
Now, as explained above, the discriminator is the maximizer and hence it tries to 
maximize

Understanding Generative Adverserial Networks - Part 1

This is a two part series on understanding Generative Adversarial Networks (GANs). This part deals with the conceptual understanding of GANs. In the second part we will try to understand the mathematics behind GANs.

Generative networks have been in use for quite a while now. And so have discriminative networks. But only in 2014 did someone get the brilliant idea of using them together. These are the generative adversarial networks. This kind of deep learning model was invented by Ian Goodfellow. When we work with data already labelled, it’s called supervised learning. It’s much easier compared to unsupervised learning, which has no predefined labels, making the task more vague. 

"Generative Adversarial Networks is the most interesting idea in the last ten years in Machine Learning." - Yann LeCun

In this post, we’ll discuss what GANs are and how they work, at a higher , more abstract level. Since 2014, many variations of the traditional GAN have come out, but the underlying conc…