What can do Metabolizer for you

Metabolizer integrates three different pathway tools:

• Activity allows you to see how module activity changes in different conditions.
• Knockout allows you to simulate knockouts or over-expressions of one or several genes or the effect of drugs in metabolic module genes.
• Prediction allows you to train a prediction model and test it with different data.

Metabolizer is able to integrate RNA-Seq and microarray data to produce an accurate result.

Metabolizer-Activity allows you to compute activity (overall flux) of each metabolic module and each one of your data samples. Therefore, it allows you to compare the activity values:

• Between two groups, in order to see which modules are changed and how

In order to check how to use these options please see Activity.

Metabolizer-Knockout allows you to modify the expression value of a set of genes and then check the effect of this change at the level of module activity. In this way, you can simulate effect of knockouts (KO) and over-expressions of one or several genes or the effect of drug(s) in the activity of metabolic modules. The simulations can be done:

• Using single sample: This option calculates fold-change of module activity between after and before KO/OE.
• Using multiple samples: This option uses one selected sample to calculate effect of in-silico intervention of gene expression on module activity. Rest of the samples are used to train a prediction model. Metabolizer uses this trained model to calculate class probability of selected sample including before and after intervention.
• User can use the options above by selecting single/multiple genes or drug(s) on interactive pathway panel. Metabolizer also allows to auto-KOs. Auto-KOs option calculates effect of simple KOs using all module genes by one by.

In order to check how to use these options please see Knockout.

Metabolizer allows you to train, download and test a prediction model for your dataset using different machine learning algorithms. Prediction can be done using following methods; SVM, Random Forest. The model can be trained and test:

• Using different groups (2 and more groups) of samples.

In order to check how to use these options please see Prediction.