Meta-analysis of shotgun sequencing of gut microbiota in Parkinson’s disease

[SvetlanaUP]

Meta-analysis of shotgun sequencing of gut microbiota in Parkinson’s disease – Hiroshi Nishiwaki – npj Parkinson's Disease
https://www.nature.com/articles/s41531-024-00724-z

[KateRasheed]

I would like to be assigned this paper please. Thank you.

[KateRasheed]

Good morning @SvetlanaUP . This paper is ready for review. Thank you.

[MyleeeA]

Can I review this @SvetlanaUP

[SvetlanaUP]

Hi @MyleeeA how is going reviewing of https://bugsigdb.org/38773112? Thanks!

[MyleeeA]

Hi @SvetlanaUP

I have been trying to be sure what was really done in the study.

I’ll leave a review now.

[MyleeeA]

Hi @KateRasheed

Well done on your curation 👏🏾, that looks like a lot of work with those signatures. I have a few observations though;

• I am of the opinion that Supplementary table 8 was not rightly curated. The table should have been curated based on Confounding factors and that would give us a total of six experiments from that table. This is because together, some of the species are increased and decreased. But when separated based on confounding factors, you have 2 directions (negative and positive).
  I’m suggesting the group name to be e.g Parkinson’s disease patients (Age adjusted) vs Healthy Controls (age adjusted), Parkinson’s disease patients (BMI adjusted) vs Healthy Controls (BMI adjusted) etc. Then you can state in the definition that the groups were adjusted for confounding factors. What’s your take on this?

• Parkinson's disease is the condition in the study and wasn’t controlled for as a confounder, please remove it.

• From this statement, though appears incomplete, I believe the location for the experiments in S8 isn’t all five locations used in the study. “ We also adjusted for confounding factors in our and USA datasets”.
  Also see this, “The metadata were available for our dataset, the USA dataset, and the German dataset[12]. _Six features (age, BMI, constipation, disease state, sex, and anticholesterol drugs), which were available for both the Japan and USA datasets but not for the German dataset, were set to possible confounding factors._ Confounding factors in the two datasets were analyzed with CPLM (Compound Poisson Linear Model) in MaAsLin2 on R 4.2.3 without any normalization or transformation for each of 73 EC numbers, 31 genera, and 38 species, that were significantly different between controls and PD in meta-analysis. The difference between the USA and Japan was set to a random effect in the MaAsLin2 analysis”. Although the title of the table suggests otherwise.

• Finally the statistical test for the experiments in Table 8 appears to be MaAslin2 (It’s found within what I copied above).

Thank you, anticipating your feedback.

[KateRasheed]

Good afternoon @MyleeeA. Thanks for the review. For Supplementary table 8, I only curated taxa with significance (illustrated with * beside its adjusted p-value). Please check this again.

Thank you.

[MyleeeA]

Thank you for pointing me towards it

Somehow I missed the fact that the adjusted p value was specifically on just Disease.