Okay, as the of my effort for Paul’s project, I am looking into this software. ChromHMM.

There was a meeting called by Dr. Paul Wade, his goal was to “associate a chromatin state to its regulated gene (expression level) changes”.

Here are two papers Ernst’s Nature paper in 2011 and Ernst’s Nature biotech in 2010 are where we could start off with.

In Paul’s mind, a pipeline will be an ideal situation. To build such a pipeline, I can think about the following component

1. Ernst's HMM model
2. Further statistical model

To learn this, we should clear out the following road block

1. Can we reproduce what these two paper proposed?
2. Can we take our data as input? Do we need any data cleaning?

These being said, to get a pipeline, we should modularize the processed into a viable components.

1. Data cleaning module
2. Chromatin state identification module
3. RNAseq/MicroArray gene expression module
4. Association module