Count dismatch between single kaiju and kaiju-multi

[jjuhh]

Hi, Kaiju team.

Using a "kaiju-multi" has the same number of lines(reads) in output file as using single "kaiju"

wc -l sample.A.kaiju.out
> 1635254
wc -l sample.B.kaiju.out
> 1949596
wc -l sample.A.B.kaiju.out
> 3584850

but, the values of counts(column 3) and percentage(coulmn 2) was not same. (after kaiju2table)

fgrep -w "67082" sample.A.kaiju.out.table
> sample.A.kaiju.out.table  0.342276	6673	67082	NA;BeAn 58058 virus
fgrep -w "67082" sample.B.kaiju.out.table
>sample.B.kaiju.out.table  0.802077	13116	67082	NA;BeAn 58058 virus
fgrep -w "67082" sample.A.B.kaiju.out.table
> sample.A.B.kaiju.out.table  0.740505	26546	67082	BeAn 58058 virus

Why is this results? and if both results are not wrong, which one is better to use, kaiju or kaiju-multi, when comparing the virome of case control study?

[pmenzel]

Hi,

would it be possible to send me the zipped output files of kaiju and kaiju2table?