# SAM in OMIQ

SAM is a differential statistical analysis method which finds the most significant stratifying signatures between files organized by metadata.

## Background

SAM (Significance Analysis of Microarrays, *Tusher, Tibshirani and Chu (2001)*) is a statistical
method used for finding significant features from input data described by a response variable. SAM is versatile in the types of statistical analysis and response variables supported. For example, two class paired and unpaired, multiclass, regression to a continuous variable, and more. In OMIQ, SAM can be used on files with metadata tags and when the workflow has filters in it. Both filter abundances and median signal of features on filters can be used as input.

SAM, as the name suggests, was originally described in the context of genomics applications. However, as noted in the documentation and literature, it can be applied to other arbitrary data. As far as we know, it was first described for use with cytometry data in Bruggner 2014, where it is incorporated as a statistical method within the Citrus algorithm. In Citrus, clusters are first generated and then run through SAM with the option of analyzing the abundance or median signal value of the clusters on a chosen set of markers. In OMIQ, any set of clusters, gates, or filters in general can be run through SAM in the same fashion which is much more flexible.

## Algorithm Settings and Setup

Note the example data and workflow used is the same as in the tutorial.