# Sperm RNA sequencing datasets — public data directory

This directory contains sequencing outputs and count matrices for eight datasets used in the sperm RNA analysis project. It is organized to make the mapping/reference choice explicit and to distinguish original count tables from consistently processed matrices used for figures.

## Datasets

| Directory | Dataset |
| --- | --- |
| `1. Sperm_SHFD14_RNA002/` | Sperm RNA002 |
| `2. sRNA_PNK_RNA002/` | Small RNA sequencing |
| `3. ILM_Sperm_SHFD14/` | Illumina sperm sequencing |
| `4. FACS_WT_RNA004/` | FACS WT RNA004 |
| `5. sRNA_PNK_5SR_Chimera_RNA004/` | Small RNA sequencing (Chimera) |
| `6. Sperm_Distal_PNK_5SR_Chimera_RNA004/` | Distal sperm Chimera RNA004 (AF42; benchmark sperm) |
| `7. Sperm_SHFD14_Chimera_RNA004/` | Sperm diet Chimera RNA004 |
| `8. FACS_Chimera_test/` | FACS Chimera RNA004 |

Directories 2 and 5 are small RNA-seq runs; the other six are the datasets covered by the common annotated-count processing workflow.

## Reference and mapping conventions

Different references and mapping approaches were tested for some runs. The reference selected for the paper's sperm sequencing analysis is:

`ref/rtRNA.GRCm39.masked.fa.gz`

Its public reference directory is:

`/enovoa/public/afefilova/reference/spermseq_paper_ref/`

The selected sperm sequencing mapping uses Leszek's latest pipeline incorporating tRNA decoder and the `spermseq_paper_ref` reference.

- **If a run contains multiple mapping/reference subdirectories**, use **`spermseq_paper_reference/`** for the paper. Other subdirectories represent alternative mapping strategies and should not be mixed into the paper analysis.
- **If a run has `bam_files/`, `counts/`, and `ref/` directly under the run directory**, those are the selected results for the paper; no extra `spermseq_paper_reference/` level is needed.
- **For small RNA-seq and Illumina runs**, outputs are stored directly in their run directories, without an alternative-mapping subdirectory. Their mapping methods should not be assumed identical to the nanopore sperm pipeline.

For example, `7. Sperm_SHFD14_Chimera_RNA004/` contains both:

```
7. Sperm_SHFD14_Chimera_RNA004/
├── spermseq_paper_reference/   # selected for paper
└── more_reduced_reference/     # alternative mapping
```

Each mapping-specific directory contains its own relevant BAM files, counts, and reference information or links.

## Directory contents

The precise layout varies by run and mapping approach, but the selected output directory generally follows:

```
<run>/[spermseq_paper_reference/]
├── bam_files/                 # alignment files and available indices
├── counts/
│   ├── raw_counts/            # original count tables from the counting pipeline
│   └── processed_counts/      # filtered, annotated matrices for downstream analysis
└── ref/                       # reference files or links used for mapping
```

The bracketed mapping subdirectory is present only where multiple mapping approaches were organized separately. Individual files and optional checksum manifests may vary between runs.

### Which counts should be used?

**Use `counts/processed_counts/` to generate the paper's plots and figures** for the six datasets in the common sperm count-processing workflow. The `raw_counts/` directory preserves the original pipeline count tables for traceability and reprocessing.

The processed matrices were generated consistently across the six datasets using an R script that:

1. Imports the relevant count matrices (including the required rRNA/tRNA features).
2. Standardizes sample identifiers.
3. Removes HTSeq summary rows while retaining `__no_feature` as `unannotated`.
4. Filters low-abundance features (at least 2 counts in at least 2 samples), with defined exceptions for preserved features.
5. Assigns gene names and RNA biotypes using GENCODE vM36 or GenBank GRCm39 annotations, as appropriate to the dataset.
6. Exports annotated matrices in `.tsv` and `.csv` formats.

The processing script and a local copy of the processed count matrices are in the lab Dropbox folder:

`Lab EvaNovoa/54.sperm_diets/__Draft/Clean Results/_counts_annotated/`

Processing script: `0.InputProcess.R` (use the final six-dataset version; check the filename/version in Dropbox).

## AF42 example

For `6. Sperm_Distal_PNK_5SR_Chimera_RNA004/`, the selected mapping output is under `spermseq_paper_reference/`:

```
6. Sperm_Distal_PNK_5SR_Chimera_RNA004/
└── spermseq_paper_reference/
    ├── bam_files/
    └── counts/
        ├── raw_counts/
        └── processed_counts/
            ├── Sperm_Distal_PNK_5SR_Chimera_RNA004_annotated.tsv
            └── Sperm_Distal_PNK_5SR_Chimera_RNA004_annotated.csv
```

The AF42 raw count tables originate from the `plots/` subdirectory of the cluster pipeline output:

`/no_backup/enovoa/users/lpryszcz/cluster/rna_mods/spermseq/dorado-2.1.0.sup.slurm/AF42_spermseq/plots/`

The corresponding processed tables were generated separately by the shared R count-processing workflow. The 12 AF42 sample columns were initially labeled by barcodes (`bc_01`–`bc_12`); confirm the experimental barcode-to-sample mapping before interpreting biological groups.

## Notes for reproducibility

- Do not combine counts or BAMs from different mapping-reference subdirectories without explicitly documenting the choice.
- Keep `raw_counts/` unchanged; use `processed_counts/` for the standardized annotated matrices.
- Where `SHA256SUMS.txt` is present, it can be checked from that directory with `sha256sum -c SHA256SUMS.txt`.
- This README describes the intended organization and processing conventions; consult the actual run-specific `ref/` links and processing script for exact inputs and parameters.

## Dataset Overview and Sample Information

The following Google spreadsheet contains additional information about the sequencing datasets:

[Sequencing Dataset Spreadsheet](https://docs.google.com/spreadsheets/d/1sG04R0qX6VuAKZ1c5uIiGb0A4e8VNQFosZWRgHyaglI/edit?gid=0#gid=0)