Referee 1's Review:

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication, the manuscript has been revised in light of these comments and to address other editorial requirements.
**********

The paper describes a very interesting potential correlation between rDNA promoter methylation, rDNA expression and suicidal individuals with/without personality disorders. Technically the paper is build on solid ground - however the quantity (bisulphite data) and partially quality (controls)) of data needs some improvement. Since the issue raised in the paper is of great psychosocial and political impact the data presented should be hard wired, firm and controlled. I cannot judge this for the psychological classification in table 1 but I have the impression that the molecular data (bisulphite/expression) compiled are too scarce and have to be improved (see points 1-5 below).

It is known that only some of rDNA genes are transcriptionally active and that this might be linked to certain alleles. This certainly complicates the comparative analysis of DNA methylation and RNA expression (which rDNA genes are indeed expressed in correlation to the bisulphite data ?) . Moreover the rDNA promoter regions are known to be highly polymorphic - the authors are not discussing this problem at all. They should be more certain of whether they have some sort of allelic bias in their bisulphite and expression analysis. This has to considered in light of the relatively small number of samples analysed. It is clear how difficult it is to obtain such a number of matching samples at all - but if you have small numbers and you are analysing multi copy genes - the data have to be extremely controlled before reaching any correlative conclusion.

Major points:
1) The number of bisulphite-clones sequenced (7-10) from individual samples is far too low for a decent statistic (10 clones = 10% technical variance) - the number should be at least doubled to have a rough estimate of the overall methylation. This particularly important for further subgrouping (cluster B versus non B suicides).
2) The „significant" correlation differences in figure mainly stems from 60 sequences which are above the correlation line. The authors should specify in the supplementary figure 1 from how many bisulphites/PCRs the indivial profiles are derived and how many polymorphisms are detectable and classify the individual profiles.
3) No other tissue (blood, liver, ...) have been included as „control" tissues - this has to be done otherwise the whole correlation with suicide and personal disorders might be heavily overinterpreted.
4) The author obviously are not worried by the possibility that the primers used for bisulphite analysis might overlap with SNPs - they should perform control sequencing of WT (non-bisulphite DNA) to test this.
5) The same holds for possible CpG/TpG SNPs in individual rDNA clusters which in some individuals may be heavily over/under represented.

Minor points:
The theoretical modelling of methylation patterns in comparison to the real data is nice but does not really clarify the issue.
The argument of promoter wide /regional methylation vs site specific methylation is not really solved as data from the Grummt lab clearly show (although this is mouse) the importance of UBF binding site demethylation for transcriptional activity (which is briefly but vaguely discussed by the authors).

Overall the text is a too long - concepts and conclusions could be written in a more concise way. Arguments are sometimes a bit blurred. The discussion does not really include a critical evaluation of the data.