# Is there a convention or standard of the highest bits in a sequence logo? [duplicate]

In bioinformatics, a sequence logo is a graphical representation of the sequence conservation of nucleotides (in a strand of DNA/RNA) or amino acids (in protein sequences).

wiki says

The total height of the letters depicts the information content of the position, in bits.

and then gives this example

A sequence logo showing the most conserved bases around the initiation codon from all human mRNAs (Kozak consensus sequence). Note that the initiation codon is not drawn to scale, otherwise the letters AUG would each have a height of 2 bits.

it seems that there is a 0.1 bit letter.

is this in the context of 2bits? is 2bits the highest bits in this sequence logo? Is there a convention or standard of the highest bits in a sequence logo?

• Sequence logos are usually just visual aids, why do you need the bits? I mean, if you want to get numbers representing the frequency of a residue at a given position, then logos are probably not the right approach. – terdon Aug 1 '19 at 9:24

This is explained in the original paper [1] describing the logos (emphasis mine):

The importance of a particular position in a binding site is more clearly and consistently given by the information required to describe the pattern there [7, 8]. To choose one symbol or state from two equally likely possibilities requires one "bit" of information. For example, to communicate the result of a coin-ip to someone requires 1 bit of information because only one yes-no question needs to be answered: "Is it heads?". If a position in a binding site always contains one base (e.g. G), then we need exactly two bits of information since two yes-no questions need to be answered: "Is it A or G?" (i.e. is it a purine?) and "Is it A or C?". (If the answers to both questions are "no", it must be a T.) Furthermore, if a position contains two bases (e.g. sometimes A and sometimes G), only one question suffices since a two out of four choice is equivalent to a one out of two choice. Therefore, only one bit is needed to describe a position in a binding site that contains only purines, but two bits are needed to describe a position that always contains adenine.

So yes, 2bits is the highest value you will see on the y-axis of a DNA sequence logo (proteins are a different story since there are more possibilities). In the one you show, the scale is from 0 to 0.2, but it's the same principle.

However, you should think of logos as a useful visual aid, and not a way of getting rigorous mathematical information about a sequence. So don't put too much weight on the bit values. If you need that sort of data, calculate it from the alignment directly.

[1] Schneider TD, Stephens RM., Sequence logos: a new way to display consensus sequences., Nucleic Acids Res. 1990 Oct 25;18(20):6097-100.