Results from scRNA-seq

Upon completion of the single-cell RNA-seq workflow, several output folders are generated. These capture each step of the pipeline, from raw quality checks and alignments to gene-level quantification and downstream-ready data objects.

This section describes the content and purpose of each output folder to help you understand what was generated and how to use it. Below is the file structure if CellRanger is chosen as the aligner.

results/
├── fastqc/
├── fq_lint/
├── trimgalore/
├── cellranger/
├── counts/
├── qc/
├── scanpy/
├── multiqc/
└── pipeline_info/

The table below describes the contents of each folder -

Folder with outputs

Folder	Contents	Description
fastqc/	FastQC .html, .zip	Raw read quality reports
fq_lint/	fastq-lint .txt, .json	FASTQ integrity and formatting checks
trimgalore/	Trimmed reads and reports	Adapter removal and quality trimming
cellranger/	Count matrices and QC reports	Primary Cell Ranger output (see below)
counts/	Processed matrices (.tsv, .h5)	Unified or filtered count matrices
qc/	Doublet detection & cell filtering	DoubletFinder, Scrublet, etc.
scanpy/	.h5ad AnnData objects	Downstream-ready files for analysis
multiqc/	Aggregated QC report	Combines reports from FastQC, Trim Galore, etc.
pipeline_info/	Logs, software versions, config	For reproducibility and debugging

---

Exploring CellRanger output

Each sample under cellranger/ contains:

cellranger/
└── sample1/
    └── outs/
        ├── filtered_feature_bc_matrix/
        │   ├── matrix.mtx.gz
        │   ├── barcodes.tsv.gz
        │   └── features.tsv.gz
        ├── metrics_summary.csv
        ├── web_summary.html
        └── ... (optional: BAM files, clustering outputs)

You can:

Open web_summary.html for key stats (e.g., # of cells, reads per cell, UMI counts).

Load the filtered matrix into Scanpy or Seurat.

Use metrics_summary.csv for structured QC across samples.

---

Output Differences with Other Pipelines

• STARsolo

If you run the workflow using STARsolo, you’ll find:

star_solo/
└── sample1/
    └── Solo.out/
        └── Gene/
            ├── matrix.mtx
            ├── features.tsv
            └── barcodes.tsv

• Outputs are similar in structure to Cell Ranger.

• You can load directly into Scanpy or convert as needed.

• Kallisto + BUStools

If using Kallisto + BUStools, the structure is:

kallisto/
└── sample1/
    ├── counts_unfiltered/
    │   ├── matrix.mtx
    │   ├── features.tsv
    │   └── barcodes.tsv
    └── counts_filtered/
        ├── matrix.mtx
        ├── features.tsv
        └── barcodes.tsv

• Filtering may be done via bustools correct and bustools count.

• Compatible with standard analysis pipelines after conversion.

• SimpleAF (Alevin-Fry) + AlevinQC

With SimpleAF, the outputs live in:

alevin/
└── sample1/
    ├── quant.json
    ├── featureDump.txt
    ├── filtered_mtx/
    │   ├── matrix.mtx
    │   ├── features.tsv
    │   └── barcodes.tsv
    └── alevinqc/
        └── alevinqc_report.html

• filtered_mtx/ is analogous to the filtered Cell Ranger output.

• alevinQC_report.html provides detailed QC like gene diversity, knee plots, and barcode filtering.

---

Output Validation Checklist

Checklist

Checkpoint	Expected outcome
Read quality (FastQC)	Q30+ average per base
Adapter trimming	Minimal residual adapter content
Cell Ranger / STARsolo matrix	Non-empty filtered_feature_bc_matrix or Gene/ matrix
UMI / gene counts	Reasonable range of cells and detected genes
QC summary (MultiQC, HTML)	No major warnings