cellmix

1. First install the cellmix R package following these instructions

   1. Also, in R, do

      1. install.packages("R.matlab")

2. The input: Download a GSE using get_GSE_data. Instructions will be written later.

3. This creates a mat file that contains the following two variables. let it be <INPUT_MAT_FILE>

• 1. expression - [genes X samples]

  2. gene_symbol - [genes X 1]

4. Create a text file markers.txt that has the following format .

5. Verify that the names of the markers exist in the variable "gene_symbol"."GENE NAME" "Class name"

6. Some marker files already were set in /cortex/code/cellmix/marker_sets/*.txt

   2. 2. 1. 1. "STMN2" "Neurons"

            2. "BRUNOL4" "Neurons"

            3. "BRUNOL5" "Neurons"

            4. "BRUNOL6" "Neurons"

            5. "SYT1" "Neurons"

            6. "GFAP" "Astrocytes"

            7. "ALDH1L1" "Astrocytes"

            8. "OLIG2" "Oligodendrocytes"

            9. "MBP" "Oligodendrocytes"

7. Run the following matlab code. The file is also located in /cortex/code/cellmix/run_cellmix.m

addpath('/cortex/code/cellmix')

input_mat_file_name = <INPUT_MAT_FILE>;

output_mat_file_name = <OUTPUT_MAT_FILE>;

marker_file_name = 'markers.txt';

algorithm = 'DECONF'; %'NMF'; 'NMFforb'; 'meanProfile'; 'DSA'; 'DECONF';

run_R_script('do_cellmix', input_mat_file_name, output_mat_file_name, marker_file_name, algorithm);