Knowledge base¶
The knowledge base is how seqforge knows what a sequencing library is. It holds one entry per technology: the read layout, the barcode positions, the barcode lists. seqforge scores your files against every entry to find the one the bytes match. The entry it picks decides how your data is parsed.
You don't choose an entry; the bytes do. These pages are here so you can see what seqforge recognises, and understand the answer when it tells you what your data is.
Supported assays¶
| assay | what it is | told apart by |
|---|---|---|
| 10x Chromium 3' v3 / v3.1 | droplet single-cell, current 3' kit | 28 bp barcode read + 10x's barcode list |
| 10x Chromium 3' v2 | droplet single-cell, previous 3' kit | 26 bp barcode read (2 bp shorter than v3) |
| BD Rhapsody WTA | microwell single-cell, 3-block cell label | two fixed linkers + three barcode lists |
| SPLiT-seq | instrument-free split-pool single-cell | two fixed linkers + three round-barcode lists |
| Bulk RNA-seq | plain bulk RNA-seq, no barcodes; paired-end or single-end | the absence of a barcode read |
How the entries are organized¶
Related chemistries are grouped into a family. The 10x 3' family, for example, is one loose pattern: "a 16 bp barcode in a 26–28 bp read". It first says "this is 10x 3' something". Then it descends to the exact member (v2, v3, or v3.1) by read length and barcode list. So seqforge can recognise a family from a vague metadata mention and still pin the exact chemistry from the bytes.
Some chemistries look the same in the reads and are processed the same way, such as v3 and v3.1. These are recorded under both names, and no question is asked. Two that look alike but are processed differently are told apart by whatever can, usually the barcode list. If nothing can, seqforge asks you rather than guessing.
Adding a technology¶
Every entry is executable: it declares a chemistry precisely enough that seqforge can generate synthetic reads from it, probe them, and check it recovers what it declared. If you want to teach seqforge a new chemistry, the Adding a technology tutorial walks through it.
For the human-readable descriptions of each library structure, the Teichmann Lab's scg_lib_structs is an excellent reference, and each assay page below links to its entry there where one exists.