Questions about structure files of customized databases

[rileyjiang]

Hi, I have a question about the structure files of customized databases. According to the Readme, the first column of the structure file should be sequences ID. Does this need to be unique? How can we deal with the situation that a sequence has several different types? For example, should I construct the structure file like:

level1 level2 level3
seq1 subtype1 type1(Ni)
seq1 subtype1 type2(Co)

or

level1 level2 level3
seq1 subtype1 type1(Ni),type2(Co)

Looking forward to your reply, Thank you!

[xinehc]

level1 need to be unique, so two seq1 is not allowed. Your second construction seems fine.

[rileyjiang]

Thank you very much for your quick reply! Another question is that what's the difference between '--structure1(single component)', '--structure2(two-component)' and '--structure3 (multi-component)'. The help page does not detail this.

[xinehc]

--structure2 is for two-component systems so each component is weighted by 0.5, --structure3 by 1/3.

[rileyjiang]

What do you mean by two-component systems? For example, I cannot see the difference between 'two-component_structure.txt' and 'multi-component_structure.txt' for the default database.

two-component:

multi-component:

Does it refer to the situation that a gene has two types or subtypes? And what's the influence on the result by using --structure2/--structure3?

[xinehc]

All genes listed in the two-component.txt file will be weighted by a factor of 0.5. The structure of the three files (single, two, multi) is identical, the only difference it the weight (1, 1/2, 1/3) applied when calculating the abundance.

[rileyjiang]

Thanks, that's clear!