CIBERSORTx is an algorithm, accessible through a web portal, designed to infer the cellular composition of bulk RNA-seq or microarray data, referred to as "mixture files". This process, known as deconvolution, involves using a source called a "signature matrix", which contains known gene expression profiles of various cell types, built by CIBERSORTx from a single-cell RNA-seq dataset referred to as a "reference matrix". By comparing the mixture files to the signature matrix, CIBERSORTx can estimate the proportions of different cell types within the sample.The files included here are used for and obtained from the construction of a kidney allograft biopsy signature matrix by CIBERSORTx, and can be downloaded to infer cell fractions on kidney allograft biopsy bulk RNA-seq or microarray data.A gene expression matrix, derived from a single-cell RNA-seq dataset (Lamarthée et al. Nat comm 2023) comprised of 35,132 cells divided into 18 different cell phenotypes and originating from 16 kidney allograft biopsies, was normalized to counts-per-million (CPM) and subset to 10,000 random cells, maintaining the relative proportion of each cell type.The resulting text file, named KTB18_reference_matrix.txt, was used as input to the "Create Signature Matrix" module of CIBERSORTx. Given that a droplet-based technique (10X Chromium) was used for generating the single-cell RNA-seq dataset, the Min. Expression value was reduced to 0.1. This adjustment aims to increase the reliability of the signature matrix, as recommended by CIBERSORTx.Upon completion, the resulting text file, named KTB18_signature_matrix.txt, can be used as input to the "Inpute Cell Fractions" module of CIBERSORTx to estimate the cell proportions of mixture files. Given again that a droplet-based technique was used for generating the single-cell RNA-seq dataset, the S-mode batch correction is necessary to remove technical differences between the signature matrix and the mixture files, since they originate from different sequencing technologies. The reference matrix is included for this purpose, as this method requires it.

CIBERSORTx is an algorithm, accessible through a web portal, designed to infer the cellular composition of bulk RNA-seq or microarray data, referred to as "mixture files". This process, known as deconvolution, involves using a source called a "signature matrix", which contains known gene expression profiles of various cell types, built by CIBERSORTx from a single-cell RNA-seq dataset referred to as a "reference matrix". By comparing the mixture files to the signature matrix, CIBERSORTx can estimate the proportions of different cell types within the sample.The files included here are used for and obtained from the construction of a kidney allograft biopsy signature matrix by CIBERSORTx, and can be downloaded to infer cell fractions on kidney allograft biopsy bulk RNA-seq or microarray data.A gene expression matrix, derived from a single-cell RNA-seq dataset (Lamarthée et al. Nat comm 2023) comprised of 35,132 cells divided into 18 different cell phenotypes and originating from 16 kidney allograft biopsies, was normalized to counts-per-million (CPM) and subset to 10,000 random cells, maintaining the relative proportion of each cell type.The resulting text file, named KTB18_reference_matrix.txt, was used as input to the "Create Signature Matrix" module of CIBERSORTx. Given that a droplet-based technique (10X Chromium) was used for generating the single-cell RNA-seq dataset, the Min. Expression value was reduced to 0.1. This adjustment aims to increase the reliability of the signature matrix, as recommended by CIBERSORTx.Upon completion, the resulting text file, named KTB18_signature_matrix.txt, can be used as input to the "Inpute Cell Fractions" module of CIBERSORTx to estimate the cell proportions of mixture files. Given again that a droplet-based technique was used for generating the single-cell RNA-seq dataset, the S-mode batch correction is necessary to remove technical differences between the signature matrix and the mixture files, since they originate from different sequencing technologies. The reference matrix is included for this purpose, as this method requires it.