config-example.yaml

# Samples.tsv: path or URL to sample sheet (TSV format, columns: sample, condition, ...).
samples: samples.tsv

# Units.tsv: path or URL to sequencing unit sheet (TSV format, 2 different formats.
units: units.tsv

# input type: this variable is used to specify the input files nature ("matrix" or "fastq"). 
input_type: "fastq"

# technology:
# "10x", "Drop-seq" or "custom" if input_type is fastq.
# "10x" or "standard" if input_type is matrix.
technology: "10x"  

# technology_version: only if using 10x technology (fastq)
technology_version: "v3"

# Output and log files paths
outdir: "out"
logdir: "log"

# Graphics availability.
graphics: True

# keeping the same random seed allows the programs that use it to generate the same results
random_seed:  4848

# gene case format: uppercase (TP53), lowercase (tp53), and titlecase (Tp53)
case: "uppercase"

# If there are to many samples, R might yield an error when writing count data in tsv files, in that case set as False.
# Possible values: True/False
write_table: True

# Perform the analysis using only single samples (for merged or integrated)
single_samples: TRUE

# genomic reference sequences and annotations
ref:
    annotation: "/my/annotation.gtf"
    fasta: "/my/genome.fasta"
    idx: "/my/genome_idx" # STAR aligner index path (auto generated if not already present)

# barcodes whitelist (10x only)
whitelist: "/my/barcodes.txt"  


parameters:
    seurat_qc: # step 1: Examine your results with the most commonly used single-cell QC metrics.
        min_cells_per_gene: 1
    seurat_postqc: # step 2: implement your filters in order to get quality cells using the information gathered in step 1.
        # NOTE: leave value empty to ignore limit
        min_feat: 500 # minimum number of unique genes detected per cell
        max_feat: 2500 # maximum number of unique genes detected per cell
        min_count: null # minimum number of counts per cell
        max_count: null # maximum number of counts per cell
        mit_pct: 10 # minimum percentaje of reads mapping to the mitochondrial genome
        ribo_pct: 40 # minimum percentaje of reads mapping to the ribosomal genome
    seurat_filter: #[OPTIONAL] step 2.1: filter out cells based on positive or negative marker gene(s)
        genes: ["CD3", "CD4"] #gene or list of genes used to filter the dataset
        threshold: 5 # expression level threshold (number of counts/cell)
        filter_out: False # True: keep cells with expression higher than *threshold*. False: the opposite.
    seurat_merge: #[OPTIONAL] step 2.5: merging all samples
        enabled: True
    seurat_normalization: # step 3: normalization and generation of jackstraw and elbow plots 
        norm_type: "SCT" # normalization strategy ("SCT" or "standard")
        variable_features: True # True: PCA calculated only using highly variable genes. False: PCA calculated using all genes. NOTE: in both cases, variable feature are calculated and stored in the seurat object. 
        regress_out:
            enabled: True # if True, regress out biological or technical effects stored in the metadata table of the Seurat object
            vars_to_regress: "percent.mt" # vars_to_regress: some variables can be regressed out to minimize their effect on the samples. It needs to be written in a python list format (more than one variable).
        regress_cell_cycle: False # if True, regress out the cell cycle effect (check out the cell cycle plot first)
        regress_merge_effect: False # if True, regress out the differences between the merged samples.
    seurat_integration: #[OPTIONAL] step 3B: integration (equivalent to the step 3 (normalization), but creates an integrated object)
        enabled: True
    seurat_find_clusters: # step 4: clustering of your data
        principal_components: 10 # this number should be set based on the output of step 3 (normalization)
        resolutions: [0.2, 0.4, 0.8, 1.2, 1.6] # the resolution value(s) will set the granularity of the clustering (typicaly between 0.4-1.2).
        k_neighbors: 20 # number of neighbors (k) used to generate the SNN graph
        batch_metadata: "condition" # Computes the Local Inverse Simpson’s Index (LISI) for all variables of interest. Specially recommended for measuring integration accuracy.
    seurat_degs: # [OPTIONAL] step 5: marker and differentially expressed genes calculation
        enabled: True
        selected_cond: [0.2, "condition"] # resolution or condition to be used in the downstream analysis. Resolutions can be written as follows: 0.2, 0.4... Conditions must be character vectors.
        ranking: True # gene ranking for all the genes to perform as GSEA
        test: "wilcox" # statistical test to use for the DE analysis (see Seurat's FindMarkers function options)
    seurat_gs: #[OPTIONAL] step 6: gene set scoring will calculate the average expression levels of each specified regulatory program
        enabled: True
        geneset_collection: "/my/symbols.gmt" # geneset collections in GMT format to test
        geneset_percentage: 0.2 # minimum ratio (expressed genes / total genes) for a geneset to be tested
    slingshot: #[OPTIONAL] step 7: trajectory inference step
        enabled: True
        selected_res: 0.4 # resolution (number of clusters) taken from Seurat object to perform the trajectory analysis.
        start_clus: False # cluster or clusters from where the curve or lineage start (optional)
        end_clus: False # cluster or clusters where the curve or lineage end (optional)
        n_var_genes: 1000 # top n variable genes selected to fit the model
        n_plotted_genes: 50 # n top variable genes to plot in the heatmap
    vision: #[OPTIONAL] step 8: cell and cluster functional analysis based on molecular signatures
        enabled: True
        geneset_collection: "/my/symbols.gmt" # geneset collections in GMT format to test
        meta_columns: ["nCount_RNA", "nFeature_RNA", "percent.mt", "percent.ribo"] # metadata information to keep from the Seurat object
        selected_res: 0.4 # resolution at which to select clusters to be added as metadata
    velocyto: #[OPTIONAL] step 9: RNA velocity plots over single experiments samples.
        enabled: True
        selected_res: 0.4 # resolution (number of clusters) to represent in the velocity plot
        downsampling:
            enabled: False # cell downsampling (useful for samples with large number of cells)
            n_cells: null # number of cells used to perform the downsampling
    fastq_screen:
        enabled: False
    fastq_screen_indexes:
        outdir: "res"
    star: # these parameters are auto selected based on the technology and version specified above. Modify only if you know what you're doing.
        10x:
            v1:
                "--soloType Droplet --soloFeatures Gene Velocyto --outFilterMultimapNmax 50 --winAnchorMultimapNmax 50 --alignEndsType EndToEnd --outReadsUnmapped Fastx --soloUMIlen 10  --soloCBlen 14 --outSAMtype BAM SortedByCoordinate --outSAMattributes NH HI nM AS CR UR CB UB GX GN sS sQ sM"
            v2:
                "--soloType Droplet --soloFeatures Gene Velocyto --outFilterMultimapNmax 50 --winAnchorMultimapNmax 50 --alignEndsType EndToEnd --outReadsUnmapped Fastx --soloUMIlen 10 --outSAMtype BAM SortedByCoordinate --outSAMattributes NH HI nM AS CR UR CB UB GX GN sS sQ sM"
            v3:
                "--soloType Droplet --soloFeatures Gene Velocyto --outFilterMultimapNmax 50 --winAnchorMultimapNmax 50 --alignEndsType EndToEnd --outReadsUnmapped Fastx --soloUMIlen 12 --outSAMtype BAM SortedByCoordinate --outSAMattributes NH HI nM AS CR UR CB UB GX GN sS sQ sM"
        drop-seq:
            "--soloType Droplet --soloFeatures Gene Velocyto --outFilterMultimapNmax 50 --winAnchorMultimapNmax 50 --alignEndsType EndToEnd --outReadsUnmapped Fastx --soloUMIlen 8 --soloCBlen 12 --outSAMtype BAM SortedByCoordinate --outSAMattributes NH HI nM AS CR UR CB UB GX GN sS sQ sM" 
        custom:
            "--soloType Droplet --soloFeatures Gene Velocyto --outFilterMultimapNmax 50 --winAnchorMultimapNmax 50 --alignEndsType EndToEnd --outReadsUnmapped Fastx --soloUMIlen N --soloCBlen N --outSAMtype BAM SortedByCoordinate --outSAMattributes NH HI nM AS CR UR CB UB GX GN sS sQ sM" 
    multiqc: "--config res/config/multiqc_config.yaml"

resources:
    default:
        threads: 1
        mem_mb: 8000
        walltime: 480
    merge:
        mem_mb: 32000
    star_index:
        threads: 8
        mem_mb: 64000
    star:
        threads: 8
        mem_mb: 64000
    fastqc:
        threads: 4
    rseqc_junction_saturation:
        mem_mb: 8000
    rseqc_readdup:
        mem_mb: 24000
    seurat_qc:
        mem_mb: 3000
    seurat_postqc:
        mem_mb: 4000
    seurat_filter:
        mem_mb: 4000
    seurat_merge:
        mem_mb: 16000
    seurat_normalization:
        threads: 4
        mem_mb: 16000
    seurat_integration:
        threads: 4
        mem_mb: 32000
    seurat_find_clusters:
        threads: 4
        mem_mb: 8000
    seurat_degs:
        threads: 4
        mem_mb: 64000
    seurat_gs:
        mem_mb: 16000
    slingshot:
        mem_mb: 8000
    vision:
        threads: 4
        mem_mb: 16000 
    velocyto:
        mem_mb: 32000
    fastq_screen_indexes:
        threads: 8
        outdir: "res"
    fastq_screen:
        threads: 8
        disabled: true