SMLG (Statistical Machine Learning Group) Discussion Forum

Query DataSets for GSE65174 (microarray)

We are only interested in comparing these samples/mice not the NRF2 KO.

GSM1588595 Atg7flox/flox mice liver #1
GSM1588596 Atg7flox/flox mice liver #2
GSM1588597 Atg7flox/flox mice liver #3
GSM1588598 Atg7flox/flox; Alb-Cre mice liver #4
GSM1588599 Atg7flox/flox; Alb-Cre mice liver #6
GSM1588600 Atg7flox/flox; Alb-Cre mice liver #7


List of genes of interest:
MTF1, MT1, MT2,MT3, MT4
Slc30a1, slc30a4,slc30a5,slc30a6,slc30a7,slc30a9,
Slc39a1, slc39a2, slc39a3, slc39a4,slc39a5,slc39a6,slc39a7,slc39a8,slc39a9,
Slc39a10, slc39a11,slc39a13 and slc39a14.

Banjo Results (220 genes + atg7 condition)

Network graphics are in .svg format.

Experimental condition (atg7 wt vs ko) is highlighted in yellow and labeled 'Experiment' in the graph.

The zip file contains the input, setting files, logs, and outputs of all 4 runs (1h, 4h, 16h, 32h run times).

Here are the scores for the 4 runs:
runtime server score
1 hour path6 -609.7951
4 hours path4 -608.7687
16 hours path3 -608.8470
32 hours path5 -608.4319

samanthagonzales wrote:Banjo Results (220 genes + atg7 condition)

Network graphics are in .svg format.

Experimental condition (atg7 wt vs ko) is highlighted in yellow and labeled 'Experiment' in the graph.

The zip file contains the input, setting files, logs, and outputs of all 4 runs (1h, 4h, 16h, 32h run times).

Here are the scores for the 4 runs:
runtime server score
1 hour path6 -609.7951
4 hours path4 -608.7687
16 hours path3 -608.8470
32 hours path5 -608.4319

Here is the color-coded dot file. Please create a dot file and a genie file only using yellow, blue, red, and purple. Also add green genes if the genes are connected to the blue, red, and purple genes with less than three genes in between.

Here is the dot file and net file for the subgraph. I also changed the condition label from "Experiment" to "Atg7KO" for clarification.

samanthagonzales wrote:Here is the dot file and net file for the subgraph. I also changed the condition label from "Experiment" to "Atg7KO" for clarification.

Now you can upload files with .net extensions. I am attaching the mouse_liver_subset.net file here.

Reuploaded 2 net files fitted using different iss values:

mliver.fit.bayes <- bn.fit(final_subgraph, data = mliver_subset, method = "bayes", iss = 0.1)
mliver.fit.bayes <- bn.fit(final_subgraph, data = mliver_subset, method = "bayes", iss = 1)

samanthagonzales wrote:Reuploaded 2 net files fitted using different iss values:

mliver.fit.bayes <- bn.fit(final_subgraph, data = mliver_subset, method = "bayes", iss = 0.1)
mliver.fit.bayes <- bn.fit(final_subgraph, data = mliver_subset, method = "bayes", iss = 1)

Great. Moving forward, iss=1 will work. Please add MT1 and MT2 to the network, update the parameters, and upload the dot and net files.

Added Mt1 and Mt2 to the subnetwork.

samanthagonzales wrote:Added Mt1 and Mt2 to the subnetwork.

Seems the network parameters were not updated properly from the data, for example, conditional probability table of Atg7KO has 0.5 for all the cells.

Sorry about that, try this one. I included it as both .net and .xdsl.