```markdown
# Goal/Experiment:
The goal of this experiment is to provide an overview of the Genomics Research Center (GRC) data delivery structure and results files for bulk RNA sequencing (RNASeq) analysis.

# Bulk RNASeq Delivery V.4

**Tyler Stahl**  
*Genomics Research Center*  
*Version 4 - October 17, 2023*

## Abstract
This protocol will give an overview of the GRC data delivery structure and results files.

## Protocol Citation
Tyler Stahl 2023. Bulk RNASeq Delivery. protocols.io https://dx.doi.org/10.17504/protocols.io.rm7vzzx9rgx1/v4

## License
This is an open-access protocol distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

## Protocol Status
Working - We use this protocol and it's working.

## Created
October 17, 2023

## Analysis Overview
The RNASeq analysis follows these stages:
1. Pre-processing (quality control/filtering/trimming)
2. Alignment
3. Post-processing
4. Feature quantification
5. Differential expression analysis between sample groups using DESeq2.

A detailed overview of the analysis can be found in the `README.txt` and `RNASeq_methods.txt` files. Below is the workflow diagram for data processing.

![RNASeq Workflow](images/RNASeq_Workflow.png)

## Delivery Structure
The delivery email includes three download links corresponding to:
1. Raw (fastq) files
2. Aligned (bam/bigWig) files
3. Results files

### Download Links
- **RNA-Seq Analysis Results and QC:** [Download Link](https://grcweb.circ.rochester.edu/pickup/230821130441-10817/deliv_NHD13_GEO_results.tar.gz) (Checksum: 1e2e1da41926b5d45ef24c103c7a5f48e)
- **RNA-Seq Analysis Aligned Data:** [Download Link](https://grcweb.circ.rochester.edu/pickup/230821130431-10715/deliv_NHD13_GEO_align.tar.gz) (Checksum: 3c57a3de7218cae0ded7f842b9b2fdd7)
- **RNA-Seq Raw Data (for GEO submission):** [Download Link](https://grcweb.circ.rochester.edu/pickup/230821130423-10606/deliv_NHD13_GEO_raw.tar.gz) (Checksum: 59b130b3986ebc9174a75ef809db1ce2)

**Note:** These URLs will expire in 10 days.

To uncompress the delivery directory, use FREE compression software [7zip](http://www.7-zip.org).

If you are on a PC, download compression software to unzip the folders. Macs have built-in zip software.

### FASTQ, BAM, and Results Files
- **.fastq files:** Contain nucleotide and quality information generated from the Illumina Sequencer.
- **.bam files:** Store alignment data and mapping quality scores in a binary format.
- **Results folder:** Contains quality control information and results files from DESeq2.

## MultiQC Report
MultiQC aggregates QC information from multiple different analysis outputs into a single interactive report.

![MultiQC Report](images/MultiQC_Report.png)

### General Statistics
The general statistics include:
- **Fastp:** % Duplication, GC content, % PF, % Adapter
- **Star:** % aligned, M aligned
- **Feature counts:** % assigned, M assigned
- **Salmon:** % aligned, M aligned

Each section is detailed below:

#### Fastp
- **Filtering statistics:** Metrics including read quality, read length, N-Content.
- **Sequencing Quality:** Phred quality scores assigned to each base. A higher Phred score means higher confidence and lower error rate.
- **N Content:** Percentage of ambiguous or unknown bases.
- **GC content:** Proportion of guanine (G) or cytosine (C) bases in the RNA sequence.

#### STAR Alignment
- **Uniquely mapped:** Reads aligned to a single loci.
- **Mapped to multiple loci:** Reads aligned to multiple loci.
- **Mapped to too many loci:** Reads aligned to excessive locations.
- **Mapped too short:** Reads aligned to genome but fall short of the filtering metrics.
- **Unmapped: other:** Non-alignable reads.

#### Feature Counts
- **Assigned:** Reads assigned to a genomic feature (i.e., gene).
- **Unassigned: Multi Mapping:** Reads aligning to multiple genomic features.
- **Unassigned: No Features:** Reads that could not be aligned to any defined genomic features.
- **Unassigned: Ambiguity:** Reads aligning to multiple features, categorized as `Ambiguity`.

#### Salmon
- **Fragment length distribution:** Refers to the distribution of fragment lengths generated in the sample.

## DESeq2 Results
Two reports in the `deSeq2` folder:
1. Star-feature (gene-level)
2. Salmon (transcript level)

### Files in Star and Salmon Folders
- **deSeq2_counts.txt:** Raw count values.
- **deSeq2_NormCounts.txt:** Count values normalized with DESeq2's median of ratio.
- **deSeq2_rlog_NormCounts.txt:** Log of the normalized counts.

### Comparison Files
Example file: `deSeq2_NHD13_vs_WT.txt`

| A      | B             | C      | D       | E         | F       |
|--------|---------------|--------|---------|-----------|---------|
| BaseMean | log2FoldChange | stat    | pvalue  | padj      |
| Hoxa9  | 636.168       | 2.557  | 21.331  | 5.92E-101 | 8.68E-97 |
| Pbx3   | 456.879       | 3.091  | 20.932  | 2.76E-97  | 2.03E-93 |
| Pbx1   | 401.557       | -3.27  | -16.477 | 5.38E-61  | 2.63E-57 |

**Terms Definitions:**
- **BaseMean:** Average expression level across samples.
- **log2FoldChange:** Log2 fold change in gene's expression between conditions/groups.
- **stat:** Test statistic to assess significance of differential expression.
- **p-value:** Calculated using a negative binomial distribution.
- **padj:** Adjusted p-value for multiple testing using Benjamini and Hochberg method.

### EnrichR Files
EnrichR is used for gene set enrichment querying four common libraries: KEGG, GO, Wiki Pathways, and ChEA. 

Note: EnrichR is not run for salmon outputs.

Example EnrichR file:

| Database  | Term                              | Overlap | P-value | Adjusted-P-value | Combined Score | Genes                                            |
|-----------|-----------------------------------|---------|---------|------------------|----------------|--------------------------------------------------|
| GO_Biological_Process_2021 | mRNA processing (GO:0006397) | 69/190  | 5.48E-22 | 2.21E-18         | 485.4247       | PUS1,PUS3,ATRX,DDX6,ZGRF1,ABCF2,ZNF148,RPL5,PUM1 |

More info on EnrichR can be found [here](https://enrichr.maayanlab.cloud/).

## FAQ
### What are the salmon results?
Salmon uses a different alignment algorithm, mapping reads at the transcript level rather than whole gene.

### Why do I not see enrichR results?
There must be at least 50 differentially expressed genes for EnrichR.

### Why is my RNASeq data showing a weak knockdown of my gene of interest despite being validated with qRT-PCR?
Discrepancies may arise due to alignment of a non-functional transcript. Viewing aligned files in a Genome Browser may help.

### Can I remove a sample from the analysis?
It is recommended not to remove samples based on clustering alone, unless there is clear experimental reasoning.

### Can the GRC re-analyze my RNA-Seq experiment?
Use the BulkDeSeq application in HyperGen, a custom no-code genomics analytics software.

### What counts files do I use and where?
Use the `deSeq2_counts.txt` files within HyperGen BulkDeqSeq application.

For Gene Set Enrichment Analysis (GSEA), use the `deSeq2_NormCounts.txt` files. Guide to get started: [GSEA Guide](https://www.protocols.io/view/gene-set-enrichment-analysis-kqdg3x67qgq25/v1)

## Further Educational Resources
- [Genomics Biology Article](https://genomebiology.biomedcentral.com/articles/10.1186/s13059-014-0550-8)
- [DGE Workshop Lesson](https://hbctraining.github.io/DGE_workshop/lessons/04_DGE_DESeq2_analysis.html)

**endofoutput**
```