For login, upload input and design data please visit Workflow.

To use Knockout tool user can upload 1 sample or a dataset.

Using 1 sample:

1) Upload Input data.

id sample
Gene1 10.02
Gene2 9.55
Gene2 2.34

2) Select Species.
3) Fill the Job information.
4) Click the Launch Job button

Uploading only 1 sapmle does not require design data and disables auto knockout. Fold change of the module activities between conditions will be shown as results of this analysis.

Using dataset (multiple samples):

Using multiple samples lets user to calculate class probability (before and after KO/OE ) of selected sample. Probabilities are calculated using Random Forest machine learning algorithm. For unbiased learning sample sizes between classes might be balanced and sample size should be good enough. Once user upload design data, statistics of the design data used as criteria to let user do probability based analysis. This statistics and warnings appears after uploading design file. If the sample sizes are imbalanced the design file will be used only to select sample for KO/OE and only fold change of the module activities between conditions will be shown as results of this analysis.

1) Upload Input data.

id	sample1	sample2	sample3
1	0.31	0.6	0.24
2	1	0.81	0.91
3	0.7	0.9	0.3
4	0.23	0.45	0.33

2) Upload Design data.

sample1 Cancer
sample2 Cancer
sample3 Healthy

if sample sizes are balanced and each group has more than 10 samples, you will get results of the class probabilities and also you will be able to use auto knockout option.

3) Select Species.
4) If you want to knockout genes by one by and automatically, you can select this option. For each gene you will get class probability including before and after KO. Even though this option executes KOs in parallel, the total analysis time can be up to 5 min.
5) Fill the Job Information form and click the Launch Job button.