expression analysis for liver metastasis from IU, IN

networks with nodes degree cutoff 50

Inspiration

tumor cell has to acquire liver-specific aberration in order to colonize at Liver, which means by looking at DEGs, we would be able to tell whether liver metastasis is more similar to primary liver cancer or its own primary sites.

What it does

How I built it

Gathering liver metastasis tissue data from GEO and Arrayexpress, intergrate them together (normalize within and across different platforms), convert probe level into gene level, use Tetrad to construct gene expression network, then do subsequent studies (pathway analysis; DEGs analysis etc.)

Challenges I ran into

Due to the large size of the data, I got some trouble running Causal-cmd properly on clusters with correct parameters. After that, analyzing the highly connected network is also a challenge.

Accomplishments that I'm proud of

I managed to get the result using causal-cmd before bridges cluster crushed. :-)

What I learned

The principle of causal model, the assumption, the difference between causal model and correlation model. The way to properly use Tetrad and causal-cmd