Based on the comments of the reviewers and the replies thereupon, a major revision is required.

While all three reviewers indicate that the scientific significance of your study is good to excellent, the comments on presentation quality are much less positive and include ''fair'' and ''poor'' assessments. In the reply letter, the authors indicate how they consider to deal with the presentation of the various figures, but I am not yet fully convinced that these changes are sufficient to take away the reviewers' concerns. The replies are on the various details of individual figures, whereas there seems to be an issue as well with the story to be told in general. I therefore strongly recommend that the authors consider which story they want to present and which thread of the story line fits best to that and according to that, which figures fit. It is currently for instance very worrying that each reviewer has a different interpretation on which dataset is the best performing dataset. This strongly suggests that presentation quality should be improved substantially.
A final note: as indicated by both authors and reviewer(s) the relationships do not have to be linear, but then why analysing them according to correlations to assume linearity?