example-causal-datasets icon indicating copy to clipboard operation
example-causal-datasets copied to clipboard
verified publisher icon cmu-phil

Ground truth problem in Sachs

Open cruyffturn opened this issue 2 years ago • 5 comments

Thank you for the package.

The edge from pip3 to plc should be reversed. 1.

cruyffturn avatar Jan 19 '23 21:01 cruyffturn

Also an edge from pka to raf is missing.

cruyffturn avatar Jan 19 '23 21:01 cruyffturn

@cruyffturn Oh thanks! I'll check it out.

jdramsey avatar Jan 19 '23 23:01 jdramsey

Huh, I pulled up the Sachs 2005 article and am staring once again at Figure 2; the edge is in fact pip3->plc. In Figure 3A, they show the plc->pip3 edge, but they show it in purple, which indicates that it is reversed. So it seems that one is OK.

However, pka->raf is definitely missing; I'll add it.

What do you think? @cruyffturn

jdramsey avatar Jan 24 '23 07:01 jdramsey

Also sorry for the delay; I had to find a couple of minutes to get the Sachs article in front of me.

jdramsey avatar Jan 24 '23 07:01 jdramsey

In this page, https://www.bnlearn.com/research/sachs05/index.html, the direction is suggested to be plc->pip3, I am not sure which one is correct.

chxliou avatar Sep 17 '23 16:09 chxliou