expression module

The expression module is responsible for determining gene expression levels from STAR bam and count files. For the highest accuracy the strandedness of the library preparation can be specified. By default, the module assumes the data is unstranded.

The expression of a configurable set of housekeeping genes is used to normalize the expression of the genes of interest.

Tools

This module relies on the STAR count files in combination with a set of housekeeping genes to normalize gene expression levels.

seAMLess is used to determine the cell type composition of the sample, as well as the resistance to Venetoclax. Please see Karakaslar et al. for details.

AMLmapR is used to predict the AML subtype, see Severens et al. for details.

Input

The minimal input for this module is one BAM file and one STAR count table specified in a PEP configuration file, as is shown below.

Minimal input for the expression module
sample_name	bam	count
SRR8615409	test/data/expression/SRR8615409.bam	test/data/expression/SRR8615409.ReadsPerGene.out.tab

For more accurate results, it is possible to specify the strandedness of your RNA library prep (unstranded, forward and reverse). If strandedness is not specified, all samples will be treated as unstranded.

Sample configuration with strandedness
sample_name	strandedness	bam	count
SRR8615409-forward	forward	test/data/expression/SRR8615409.bam	test/data/expression/SRR8615409.ReadsPerGene.out.tab
SRR8615409-reverse	reverse	test/data/expression/SRR8615409.bam	test/data/expression/SRR8615409.ReadsPerGene.out.tab
SRR8615409-unstranded	unstranded	test/data/expression/SRR8615409.bam	test/data/expression/SRR8615409.ReadsPerGene.out.tab

Output

The expression levels for genes specified under the report section of the configuration will be included in the HAMLET pdf report.
All genes from the bed file and the genes_of_interest will be included in the MultiQC report. Stranded (forward and reverse) and unstranded samples will be listed separately, since their values cannot be compared directly.
The cell type composition for each sample wil be included in the HAMLET pdf report. In addition, an overview of the composition of all samples is included in the MultiQC report.
The predicted AML subtypes based on AMLmapR will be included in the PDF report.

Note that the overexpression and cell type composition results can be exported to csv directly from the MultiQC html file.

Configuration

The following options are available for the expression module

Configuration options

Configuration options
Option	Description	Required
seamless_ref	Path to seAMLess reference expression csv file (genes x samples)	yes
seamless_meta	Path to seAMLess metadata csv file (samples x features)	yes
gtf	A GTF file, to look up the ENSG for the housekeeping genes	yes
housekeeping	A list of genes to use for normalizing the expression	yes
bed	A BED file with genomic regions (genes) to quantify	no
genes_of_interest	A list of gene names to quantify (must be present in the gtf)	no
report	Genes to include in the PDF report, can include names from the bed file	no

Example

{
  "seamless_ref": "test/data/seAMLess/seAMLess_expression_reference.csv",
  "seamless_meta": "test/data/seAMLess/seAMLess_expression_metadata.csv",
  "gtf": "test/data/reference/hamlet-ref.gtf",
  "housekeeping": [
    "MT-CO2"
  ],
  "bed": "path/to/bed/file.bed",
  "genes_of_interest": [
    "MT-ND3",
    "MT-ND2"
  ],
  "report": [
    "MT-ND3"
  ],
}