Institute for Immunity, Transplantation and Infection


Now available: Excel based cell-type specific deconvolution package

Natural fluctuations in sample cell-type subset frequencies are a major hindrance to identifying disease associated differentially expressed genes in microarray experiments.  csSAM (cell-type specific Significance Analysis of Microarrays) is a suite of statistical procedures [1], aimed at maximizing the information obtainable from a heterogeneous tissue gene expression assay. Given data from heterogeneous tissue and information about the cell-type subset composition of each sample, the csSAM methodology statistically deconvolves cell-type specific expression profiles and utilizes them to adjust for heterogeneity, reconstruct expression profiles with one or more cell-types removed, identify the subset source of expression and perform cell-type-specific differential expression analysis. 

Thus, as an approach, csSAM’s potential to alter current practices in genomic data analyses and yield enhanced signal and biological understanding are great and extend beyond human immunology. However, its wide spread use to date has been limited due to its implementation as a set of functions in R, and hence requiring a strong familiarity with computer programming to operate.  Based on feedback from the immunology community, we have developed a version of csSAM which can be operated from within Microsoft Office Excel as an Add-In, effectively removing the requirement for user computer programming skills. The Excel version of csSAM requires that the user input the sample gene expression data on one spreadsheet and the matching cell-type proportions on another. A dialog box allows tuning of different parameters and csSAM execution occurs with a click of a single button. Simultaneously, we have rewritten the original R code to be more modular so that it may be more easily integrated with other programmatic based analyses. The output of the Excel package matches that of the original R code.  The package is available freely to academic users upon signing a software agreement and can be licensed to commercial users by Stanford’s Office of Technology And Licensing. More information is available here.


Stanford Medicine Resources:

Footer Links: