Probabilistic Graphical Models, Fall 2017

[cwyoo]

Dynamic Bayesian Networks Example

DBN-example.jpg (459.33 KiB) Viewed 5166 times

Dynamic Bayesian Networks Example

[jumorale]

I have created my table which contains 7 microarray datasets from GEO Datasets.
I will try to categorize them by the levels of chemical exposure
My species of interest is (Danio Rerio) fish species= zebrafish
I have collected a total of 100 samples
To continue I will await to run the R script to compare manual and automated methods

[Kaumudi]

This week I have been focusing on generating my input gene-list for comparing the two patient groups. I am using key transcription factors in cancer signaling pathways identified from the literature in order to narrow down my input genes for Banjo.

[musfiqsazal]

Hello Dr. Yoo and others,

I am looking for some open source code for Bayesian Networks creation, visualization and learning. I prefer R, but Python is also okay. If you know about any good source, please let me know and share the link.

Thanks,
Musfiqur

[Kaumudi]

week of 10/23/17 - worked on getting data ready to input for Banjo analysis

[jumorale]

This week I cleaned up the data since it will soon be ready for BANJO.
I have a total of 997 matched genes and a total of 100 samples.
I am in the process of discretizing my data to 0,1,2

[musfiqsazal]

Update: I am going to implement my project in R. I have collected and cleaned up data from Human Microbiome Project (HMP). The bacterium data (oral microbiome) are from seven sites of the mouth. I have normalized the abundance data and extract abundance matrix for each microbe data. Next step is to create Bayesian Networks from data, for creation I will follow learning from data approach.

[Kaumudi]

Started running Banjo this week

[Kaumudi]

Based on feedback from Dr. Yoo at our weekly group meeting, I revisited my data discretization strategy. I also spent some time refining my understanding about the exact way in which raw data is converted to Z-scores.

[musfiqsazal]

Weekly Update:
I have created a Bayesian Networks by learning from data method using R. At first, I didn't use all the data, I used a sample data to create that initial BNs. Now I will check the goodness of the model and if it is a good model, then I will go further.