Differences in KEGG reads total

User Questions
1

Dear MEGAN community,

I am a new user for MEGAN. I ran diamond to generate ‘daa’ file. Then meganised using daa-meganizer command.

I want to summarize the reads counts by KEGG groups.
But the “Reads total” was no match to sum of KEGG groups. (e.g. KEGG[365497] =/= Metabolism[25290] + Genetic Information Processing[10624] … + Not assigned[290313])

In my understanding, each KEGG ID (e.g K19244 alanine dehydrogenase [EC:1.4.1.1]) belongs to the KEGG module (e.g ko00430). And the KEGG groups contain the KEGG modules.

I attached the List Summary.
The sample that I used, contains 1,105,227 reads.

########## Begin of summary for file: /data1/nuri.megan/diamond.daa/1-Day0_AL_1.daa
Samples: 1
Reads total: 365497
Summarized at nodes:
KEGG: 365497.0
Metabolism: 25290.0
Genetic Information Processing: 10624.0
Environmental Information Processing: 8761.0
Cellular Processes: 7685.0
Organismal Systems: 4965.0
Human Diseases: 7085.0
Brite Hierarchies: 45889.0
Not Included in Pathway or Brite: 9331.0
Not assigned: 290313.0
########## End of summary for file: /data1/nuri.megan/diamond.daa/1-Day0_AL_1.daa

Many thanks in advance.

Kind regards,
NuRi

2

Each read is assigned to at most one KO and each KO appears as a leave in the KEGG tree. However, it is important to note that the same KO may appear as a leaf in multiple different locations of the tree. So, potentially, a read can be counted more than once.

For a given internal node, MEGAN reports the number of different reads that occur below it, counting each read only once (even if its KO appears in multiple different places below the internal node).

So, if N(v) is the number of reads reported for a given node v, and if v has children w1, w2, w3, … wk, then in general N(v) < N(w1) + N(w2) + N(w3)+… N(wk), because
each read will be counted only once in N(v), and only once in N(w1) and once in N(w2) etc, but may well be counted multiple times in N(w1) + N(w2) + N(w3)+… N(wk), for example, once in N(w1) and once in N(w2)…

Was that your question?