> ## Documentation Index
> Fetch the complete documentation index at: https://docs.catalog.igvf.org/llms.txt
> Use this file to discover all available pages before exploring further.

# Files & Filesets

> How files_fileset links Catalog data to IGVF and ENCODE portal files

# Files & Filesets (`files_fileset`)

The Catalog **`files_fileset`** collection connects knowledge-graph edges to the underlying **source files** on the [IGVF Data Portal](https://data.igvf.org) or [ENCODE Portal](https://encodeproject.org).

## Accession patterns

| Pattern              | Portal            | Example                           |
| -------------------- | ----------------- | --------------------------------- |
| `IGVFFI…`            | IGVF              | `IGVFFI9602ILPC`                  |
| `ENCFF…`             | ENCODE            | `ENCFF968BZL`                     |
| `ENCSR…` / `IGVFDS…` | Fileset / dataset | Parent experiment or analysis set |

## Portal metadata

Each file record on the portal includes:

* **`file_format`** — e.g. `tsv`, `bed`, `bigBed`
* **`output_type`** — processed data category
* **`file_format_specification`** — link to the document defining **column names and types** in the file (critical for tracing fields through adapters)
* Assay, lab, biosample, and release status

The Catalog loading pipeline reads these portal objects in [`file_fileset_adapter.py`](https://github.com/IGVF-DACC/igvf-catalog/blob/dev/data/adapters/file_fileset_adapter.py) when building `files_filesets` nodes.

## Querying by fileset

Use `files_fileset=<accession>` on many endpoints to restrict results to a single dataset, for example:

```
GET /variants/genes?files_fileset=IGVFFI9602ILPC&page=0
GET /files-filesets?file_fileset_id=ENCFF004PFU&page=0
```

## Field lineage

Column names in the raw portal file are transformed by **KG loading adapters** before appearing in API responses. See [Field lineage](/data-sources/field-lineage/index) for portal → adapter → Arango → API mappings.

## Browse all indexed files

See the [Data Sources index](/data-sources/index) for a table of file accessions referenced in this documentation with portal and format-spec links.