More docs on the ARB website.
See also index of helppages.
Last update on 01. Sep 2022 .
Main topics:
Related topics:

The integrated aligners


ARB_EDIT4/Edit/Integrated Aligners



Currently there are two integrated aligners:

  1. Fast Aligner
  2. Island Hopper (see Subtopic)

The following adjustments and features should apply to both aligners.

We did not test everything yet with island hopper, so some of them are broken. Please mail to if you find something.




Align current, marked or selected sequences.
If you type 'CTRL-A' in the main editor window this option is set to align the current species and the aligner gets called.


The aligner needs a sequence as reference. You can either
  • select a fixed species by name,
  • the consensus of the group containing the aligned species or
  • the next relative(s) found by the selected PT-Server.

If you choose 'Species by name', you may press the 'COPY' button to copy the name of the 'Current Species' to the 'Reference' species. Alternatively you may use CTRL-R while the focus is inside the sequence view (Note: CTRL-R does not work, if ´View differences to selected´ is active).
If you choose 'Auto search by pt_server', the aligner will use the next relative(s) as reference.
  • Please read section about 'Protein alignment with pt_server' below.
  • If the nearest relative has gaps where the sequence to align has bases, the aligner will use the 2nd nearest relative or if that one has gaps too, the 3rd nearest, etc. You can define the maximum number of relatives considered.
  • All used relatives and the number of base positions used from each relative, will be written into the field 'used_rels' (see also ´Mark by reference´).

If you enter '0' in 'Data from range only, plus', relative search only uses data from the aligned range. If you enter a value different from '0' the used range will be expanded (positive values) or limited (negative values). When the input field is empty, the complete sequence will be used.
Press 'More settings' to define how relative search works in detail. See ´Nearest relative search´


Align only a part of or the whole sequence.
Several possibilities exist for aligning just a part of the sequence:
  • select 'Positions around cursor' and specify how many positions shall be taken into each direction from the cursor position (Example: If you align 10 columns around position 100 then columns 90-110 will be aligned).
  • if you use 'Selected range' the column range of the selected block will be used.
  • if you select 'Multi-Range by SAI', the specified SAI will be interpreted as a list of ranges. A list of characters defines what is considered a range. All ranges will be aligned.

See also ´Modify SAI range´ for howto create suitable SAIs.

Turn check

The aligner is able to detect sequences which were entered in the wrong direction. With this switch you can select, if you like the aligner to turn such sequences and if it should ask you.
NOTE: In two cases turn checking isn't reasonable:
If you align only a part of a sequence or if you do not search Reference via pt_server. In both cases turn checking will be disabled.


The aligner can generate reports for the aligned sequence and for the reference sequence. These reports can be viewed with EDIT4, if you choose File/Load Configuration/DEFAULT_CONFIGURATION
The report for the reference sequence (AMI) contains a '>' for every position were the aligner needed an insert in the reference sequence.
The report for the aligned sequence (ASC) contains the following characters:
'-' for matching positions
'+' for inserts (in aligned sequence and in reference sequence)
'~' for matching, but not equal bases (A aligned to G, C aligned to T or U)
'#' for mismatching positions

Protein alignment with pt_server  

If you want to align protein sequences and use a PT-Server (to detect the next relative for each sequence), you need to

  • have two alignments in your database (a protein alignment and a corresponding DNA alignment). ARB has functions to synchronize these alignments (see ´Recommended way to maintain amino acid alignments´),
  • build a pt_server based on the DNA-alignment, select that pt_server in the aligner window and
  • specify the name of the DNA-alignment in the 'Alignment' field.



This aligner knows about and uses all extended base characters (ACGTUMRWSYKVHDN) for the alignment. In other words: M aligned to R costs no penalty.

The config-manager icon handles the settings in the 'Integrated Aligners' window and those in its subwindows 'Parameters for Island Hopping' and 'Family search parameters'.









If you select the menu entry 'remove all aligner entries' ARB_EDIT4 crashes in most cases.


  1. Close all groups containing species with aligner entries, so that no aligner entries are visible.
  2. Remove all aligner entries
  3. Reload configuration