SILVA databanks for FROGS
=========================

Source dataset
--------------
All SILVA databanks distributed for FROGS are built from the SILVA "Ref NR 99"
datasets (https://www.arb-silva.de/):
  - SSU Ref NR 99 for the 16S and 18S databanks
  - LSU Ref NR 99 for the 23S and 28S databanks

"Ref" is the high-quality, nearly full-length subset of SILVA "Parc".
"NR 99" means the Ref sequences have been clustered at 99% identity, keeping
one representative sequence per cluster (sequences from cultivated species
are always retained).

This has been the case for all SILVA releases distributed with FROGS.

Amplicon-specific databanks
---------------------------
SILVA SSU and LSU datasets contain all three domains of life. The
amplicon-specific databanks are obtained by domain extraction:
  - 16S = SSU, Bacteria + Archaea
  - 18S = SSU, Eukaryota
  - 23S = LSU, Bacteria + Archaea
  - 28S = LSU, Eukaryota

Changes from SILVA release 144
------------------------------
1. Truncated sequences ("trunc" files)
   From release 144 onwards, FROGS databanks are built from the SILVA
   "*_tax_silva_trunc.fasta" files. In these files, all nucleotides that do
   not align to the SILVA rRNA reference alignment have been removed (e.g.
   flanking ITS/spacer regions, primer or vector remnants), so that only the
   rRNA gene region is retained.
   Previous releases were built from the non-truncated
   "*_tax_silva.fasta" files, which contain the full region as deposited in
   ENA/GenBank/DDBJ.

2. No Pintail-filtered databanks
   SILVA discontinued the Pintail quality score in release 144. Databanks
   filtered on a Pintail threshold (e.g. "16S_pintail50", "16S_pintail80",
   "16S_pintail100") can therefore no longer be provided. From release 144
   onwards, only unfiltered databanks are distributed.