---
license: mit
---

# PiperNET data

Data for PiperNET, a multi-*omics* platform for elucidating the biosynthetic origin of plant specialized metabolites.
It holds LC-MS metabolomics and RNA-Seq transcriptomics data from several *Piper* species and tissues, with biological replicates.
Code and pipeline: [github.com/titodamiani/PiperNET](https://github.com/titodamiani/PiperNET).

## Download
With the [`hf` command line tool](https://huggingface.co/docs/huggingface_hub/guides/cli):
```bash
hf download titodamiani/PiperNET --repo-type dataset --local-dir data
```
To download one folder only, add `--include`, for example `--include "processed/*"`.
To run the code on this data, follow the installation steps in the [code repository](https://github.com/titodamiani/PiperNET).

## Structure
```
data/
├── README.md                        # this dataset card
├── raw/
│   ├── lcms/
│   │   ├── rawfiles/                # LC-MS raw files, mzML (140 files)
│   │   └── metadata.tsv             # sample metadata, read by MZmine
│   └── rnaseq/piperNN/              # transXpress output, one folder per sample ID (13)
│       ├── transcriptome.fasta      # de novo assembly
│       ├── transcriptome.pep        # predicted proteins
│       ├── transcriptome_expression_isoform.tsv
│       ├── busco_report.txt
│       └── annotations/             # Pfam, BLASTp, SignalP, TargetP, TMHMM
├── interim/
│   ├── lcms_mzmine/                 # MZmine output: feature table, MS/MS spectra (.mgf), annotations, networks (.graphml)
│   └── rnaseq_clstr/piperNN/        # CD-HIT clustered proteomes
├── external/
│   ├── molecular_networks/
│   │   └── gnps2/                   # GNPS2 feature-based molecular networking results
│   ├── sirius/                      # SIRIUS results: CSI:FingerID structures, CANOPUS classes
│   ├── orthogroups/
│   │   ├── sonicpd/                 # SonicParanoid orthogroups and the genes in no orthogroup
│   │   └── orthofinder/             # OrthoFinder orthogroups, MCL clusters, species tree and run log
│   ├── speclibs/                    # MS/MS spectral libraries for annotation in MZmine
│   ├── known_enzymes/               # known enzymes table
│   ├── customDB.csv                 # manually curated list of known LC-MS features, for targeted feature detection in MZmine
│   ├── standardDB.csv               # read by MZmine and the LC-MS data preparation
│   └── novelty_scores.csv           # preliminary
└── processed/
    ├── ftable_clean.csv             # LC-MS feature table
    ├── ntable_clean.csv             # GNPS2 node table, for Cytoscape
    ├── proteomes/piperNN/           # proteome.pep, proteome.csv, blastDB/
    └── proteomes_all.csv            # all proteomes with SonicParanoid and OrthoFinder orthogroups
```

## IDs
- `ftable_clean.csv` and `ntable_clean.csv`: `network_id` > 0 is a molecular network. A singleton feature is its own network, with `network_id` = -`feat_id`.
- `proteomes_all.csv`: 4 columns per orthogroup tool, with the prefix `sonicpd_` (SonicParanoid) or `orthofinder_` (OrthoFinder): `_id`, `_size` (genes), `_sp_in_grp` (species), `_intersp` (species joined with `;`).
- A singleton transcript is its own orthogroup, with a negative ID (-1, -2, ...). `sonicpd_id` starts at 1. `orthofinder_id` is the number in the OrthoFinder ID (`OG0000123` is 123), so it starts at 0.

`scripts/README.md` in the [code repository](https://github.com/titodamiani/PiperNET) describes how each file is made.