Column Statistic (Prototype)

ARB_EDIT4/View/Activate column statistics

ARB_EDIT4/View/Disable column statistics

ARB_EDIT4/View/Toggle detailed Col.-Stat.

ARB_EDIT4/View/Set threshold for D.c.s.

Highlights unlikely bases in an alignment using the maximum likelihood technique.

As soon as the go button is pressed, the tree selected in ARB_NT and all marked sequences are read. Then the relative likelihood for each base is calculated and transformed to a number between 0 and 9. This number is translated into a color, using the colors 'RANGE 0 ... RANGE 9'. The higher the number, the more unlikely a particular base.

You may improve the output by selecting a valid column statistic, which holds information about column dependent rates and base frequencies.

To change the colors used for column statistics, edit "RANGE 0" up to "RANGE 9" in ´Color properties´.

All sequences that should be analyzed, have to be marked and have to be members of the tree currently shown in the ARB main window! The program assumes that your tree has good quality - if it does not, results may be quite useless.

See also ´Sequence background color priority´.

This is only a prototype, don't expect something perfect.

Without a powerful computer only a small number of sequences can be viewed.

The program can only be started once.

This special mode displays 4 rows below the sequence containing the likelihood of each base character.

Each row consists of two rows of digits displayed in the same color (which actually is the color normally used for displaying the appropriate base character). The upper of these two rows is the first, the lower the second digit of the likelihood. [You may like to use the cursor to simplify reading]

Special character used:
-----------------------
SPACE   = likelihood is 20%
^       = likelihood is 100%
?       = can't determine likelihood for that column

Background colors used: ----------------------- Normally the normal background color is used.

If a column has a significant likelihood for one base character (or for the sum of two base characters), all four rows are displayed in color 'Range 0'.

[significance is 90% (hardcoded) - will be made utilizable soon]

The single (or two) base character(s) responsive for the significance will be displayed in colors 'Range 1' to 'Range 8' (the higher the Range-number is, the higher is the likelihood)