Inputs & Outputs

SCALT.py requires a scRNA seq row counts matrix as input data. The matrix must be in .tsv extension reporting:

1. genes on the rows;

2. cells on the columns.

The first row of the matrix must contain the ids of each cell while the first column must provide the gene ids written either as gene symbol or ensembl id (not together).

Note

In the case of enesembl gene id, the usage of the version of the id is not allowed. For example, ENSG00000235430.1 cannot be used in the counts matrix. Use ENSG00000235430 instead.

An example of the input file is reported below.

genes/cells	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20
ENSG00000160072	0	0	1	5	7	0	0	1	9	0	10	1	0	11	2	0	0	0	1	1
ENSG00000142611	4	10	0	0	0	0	1	3	1	12	10	1	0	9	7	0	0	0	0	4
ENSG00000157911	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
ENSG00000142655	0	0	0	0	16	14	0	0	0	0	0	0	0	20	0	0	0	0	0	0
ENSG00000149527	1	1	3	2	1	2	0	1	0	5	1	1	0	2	3	1	0	5	1	1

The application returns two outputs:

1. a report file in .html format;

2. a directory named results_directory hosting a collection of metadata produced upon classification.

The metadata directory lists a series of files which are produced automatically during the classification step and are required for the generation of the report. Among them, we find:

1. p_values.tsv, a tabular file reporting a collection of p-values per each cell. Each p-value indicates to a likehood test. Therefore, the number corresponds to the number of cell types tested;

2. deltas.tsv, a tabular file reporting a series of likelihood differences between the cell type specific model and the mean cell type. The number corresponds to the number of cell types tested;

3. originalTables_zipped.zip, a zipped file containing the original counts;

4. _adj.tsv file. Counts table after input set-up performed by SCALT.py by default;

5. _adj_genesExpressed_filter.tsv file. Tabular file reporting either PASS or EXCLUDE if the cell expresses at least the minimum number of genes set in the -Min (or –Threshold) parameter or not;

6. barplot_cellTypesAboundance.html, barplot_survivedCells.html, UMAP_2D.html and UMAP_3D.html. Collection of plots visualized in the report file.

SCALT parameters

SCALT.py makes usage of a collection of both positional arguments and parameters that can visualized typing the following command:

python3 SCALT.py -h

Or:

python3 SCALT --help

The documentation should appear as follows:

usage: SCALT.py [-h] [-Min --Threshold] [-Notation --Notation]
              [-Types --Types] [-CPUs --CPUs]
              Sample

1. Sample is the only positional argument of the tool. It represents the scRNA seq counts matrix file;

2. -Min or –Threshold is the minimum number of genes that a cell must express to be classified. The default value is 250;

3. -Notation or –Notation is the type of gene notation present in the counts. The defaul is ensembl id. Instead, write gene_symbol to switch to the gene symbol nomenclature;

4. -Types or –Types is the name of the directory containg the lists of the cell types to use in the likelihood test. By default, only the pre-compiled lists (DISCO, HPA) are used. To use only the custom lists generated from annotation, insert custom. Finally, to utilize only the custom lists generated from the user-defined lists, insert naive;

5. -CPUs or –CPUs is number of processors employed. The default is 1;

6. -h or –help shows the documentation.

Run SCALT.py

SCALT.py is quite straightforward sine it requires just the counts table as positional input.

Leaving default parameters, the basic command appears as follows:

python3 SCALT.py read_counts.tsv

By default, the ensembl id is used.

If the gene symbol is used in the counts matrix, the notation must be specified as follows:

python3 SCALT.py read_counts.tsv -Notation gene_symbol

Or:

python3 SCALT.py read_counts.tsv --Notation gene_symbol

By default, a cell is classified if it expresses at least 250 genes. Managing the SCALT.py parameters, this threshold che be modified as follows:

python3 SCALT.py read_counts.tsv -Min 500

Or:

python3 SCALT.py read_counts.tsv --Threshold 500

In addition, the computational time can be reduced if the number of processors is increased as reported:

python3 SCALT.py read_counts.tsv -CPUs 4

Or:

python3 SCALT.py read_counts.tsv --CPUs 4

Make sure to have available the number of desidered processors in your machine.

To conclude, the different parameters can be modified in a unique call:

python3 SCALT.py read_counts.tsv -Notation gene_symbol -Min 500 -CPUs 4

Or:

python3 SCALT.py read_counts.tsv --Notation gene_symbol --Threshold 500 --CPUs 4

The order of parameters is irrelevant.

Report

The report is a file in html format composed of a collection of plots summarizing the general statistics and classification results of the analysis. The file reports four different plots:

1. a bar plot showing how many cells express or not the minimum number of genes for classification;

2. a second barplot counting how many cells were classified to a cell type cathegory;

3. a 2D UMAP;

4. a 3D UMAP.

Note

The genes used for the creation of the UMAPs coordinates are the union of genes coming from the 471 lists of genes without repetitions.

Workflow

Running SCALT.py, the following workflow is performed:

1. inputPreparation.py is a python script that adjustes the input counts table in order to be properly analyzed by SCALT.py;

2. likelihood_ratio_test.py is the python script that performs the actual likelihood test;

3. reportGenerator.py is the python script that creates the final report.