More docs on the ARB website. See also index of helppages. Last update on 04. Mar 2022 .
Main topics:
Related topics:

# How is the consensus calculated?

## DESCRIPTION

### What to do with gaps?

Define whether to use or to ignore gaps entirely:
If you count gaps and the gap frequency exceeds the specified threshold, the consensus will show a '-'.
If the switch is 'off', the algorithm will virtually remove all gaps. That means if you have a column with 10 'A's and 500 gaps the program thinks of 100% 'A' (if the switch is 'on', the relative number of 'A's would be 2%).
Regardless of how gaps are handled, if a column contains only gaps, the consensus will always show a '='.

### Simplify using base character groups

If grouping is 'off' the most frequent character will be used in consensus (even if another character has the same frequency).
If grouping is 'on' base characters will be grouped as follows:
• RNA/DNA alignments use the IUPAC ambiguity codes (MRWSYKVHDBN)
• Amino acid alignments use amino acid classes (ADHIFC)

Click the "Show IUPAC" button to display detailed information about these character groups.
The threshold defines how characters are grouped:
• in RNA/DNA alignments the threshold specifies, whether a non-ambiguous character is considered for grouping (i.e. all characters below the threshold will be removed and the rest will be grouped. See also example below)
• in amino acid alignments the threshold specifies, whether all characters of a group together are frequent enough to show that group in the consensus. If not, an 'X' will be displayed in consensus.

Example:
If you have 40% 'A', 10% 'C', 40% 'G' and 10% 'T' and 'threshold for character' is set to 20%, arb will use the iupac code representing 'A' and 'G' (i.e. 'R').
Reasonable thresholds explained:
• amino acid:
• 51% means: if most belong to one amino acid group, then show it (otherwise show 'X')

• RNA/DNA:
• 26% means: group up to 3 nucleotides in an ambiguity code (otherwise show most frequent base)
• 25% means: group up to 4 nucleotides, i.e. will produce 'N' in consensus (only if the nucleotides are distributed EXACTLY even)
• 20% is similar to 25%, but also slightly uneven distributions will produce 'N'
• 51% will effectively turn off IUPAC grouping for nucleotides
• 50% will group 2 nucleotides (only if they are distributed EXACTLY even)
• 33% will group up to 3 nucleotides (only if they are distributed EXACTLY even)

### Show as upper or lower case?

Define whether the character is displayed in upper or lower case or whether a dot is displayed.
Define upper and lower limit:
If the percentage of a character is above or equal to the upper limit, the character is displayed in upper case.
If the percentage of a character is above or equal to the lower limit and below the upper limit, the character is displayed in lower case.
Otherwise a dot ('.') is displayed.
If gaps are ignored (as explained at top), the percentage is calculated relative to all existing bases in the column. If gaps are NOT ignored, the percentage is calculated relative to the number of species.

## NOTES

### You can save/load the consensus settings to/from a file using the config-manager icon. This allows you to exchange

• the consensus settings used in EDIT4 and
• the settings used to calculate the SAI 'CONSENSUS' from the ARB main window.