Chris Oldnall

Chris Oldnall

PhD researcher with the MAC-MIGS programme.

Manuscript Revisions Plan

Manuscript Revisions Plan

This page outlines the detailed workflow plan for completing the manuscript revisions between 23rd December 2023 and 3rd January 2024. Each day is divided into tasks and expected outputs to ensure all reviewer feedback is addressed systematically. Reviewer points are listed at the bottom of the page for reference, with task numbers linked to these points.

Workflow Plan

23rd December

Tasks

  1. Review reviewer feedback to familiarise yourself with comments. (Major Point #8, Minor Point #1)
  2. Organise paper sections into areas needing revisions. (Major Point #8, Minor Point #1)
  3. Draft a rebuttal template with placeholders for each comment. (Major Point #8, Minor Point #1)

    Output

    • Rebuttal template.
    • Organised task tracking document.

24th December

Tasks

  1. Standardise SNP selection methods across PBMC and UKB datasets. (Major Point #3)
  2. Investigate and document sample size discrepancies (861 PBMC samples vs 736 individuals in GWAS). (Minor Point #4)
  3. Add clarifications to the Methods section, including genetic instrument criteria (LD ( r^2 ), p-value thresholds). (Major Point #6)
  4. Address potential sample overlap between datasets and its possible bias. (Major Point #5)

    Output

    • Updated Methods section.
    • Supplementary note or table clarifying SNP selection and overlap bias.

26th December

Tasks

  1. Define key terms like “isoform-specific proteins,” “protein group,” and “modification-specific peptides.” (Major Point #10)
  2. Create a table/figure to visually explain these terms. (Major Point #10)
  3. Ensure consistent terminology in Sections 2.1, 4.1, and captions. (Major Point #10)

    Output

    • Glossary in the Methods section.
    • Terminology table or figure.

27th December

Tasks

  1. Justify use of PBMC data in Section 4.4 and expand its rationale for inclusion in the context of CVD. (Major Point #1, Minor Point #1)
  2. Expand on how HEIDI filtering thresholds were determined and applied (Section 4.6). (Major Point #7)
  3. Add discussion on MR assumptions and address the limitations of single-SNP instruments for PBMC data. (Major Point #2)

    Output

    • Revised Methods section (HEIDI thresholds and PBMC data rationale).
    • Expanded “Study Limitations” subsection (Section 3).

28th December

Tasks

  1. Prepare a table comparing GS and UKB cohort demographics (Section 4.1). (Major Point #5)
  2. Perform sensitivity analyses to address sample overlap using non-overlapping datasets or independent consortia data. (Major Point #5)
  3. Incorporate additional MR methods (e.g., IVW, MR-Egger, RAPS) and compare results to GSMR. (Major Point #9)

    Output

    • Cohort demographics comparison table.
    • Sensitivity analysis results.
    • Validation of MR results using additional methods.

29th December

Tasks

  1. Clarify lead SNP selection criteria in Sections 2.1 and 2.3.1. (Major Point #6, Minor Point #8)
  2. Expand discussion of COMT and SWAP70 discordant effects, including biological contributions like tissue-specific functions. (Major Point #4)
  3. Address inconsistencies in protein association numbers (127 vs 165 proteins). (Minor Point #2)

    Output

    • Refined Results section (COMT, SWAP70, and SNP selection).
    • Supplementary tables for SNP selection.

30th December

Tasks

  1. Add histograms or numerical summaries for sample sizes (Section 4.1.8). (Minor Point #4)
  2. Create supplementary tables for all tested proteins with significance. (Minor Point #7)
  3. Include genome reference details for BP coordinates. (Minor Point #7)

    Output

    • Updated supplementary materials.
    • Figures for sample size variation.

31st December

Tasks

  1. Add a detailed “Study Limitations” subsection in Section 3, discussing:
    • Statistical power limitations of PBMC analyses. (Major Point #2)
    • Sample overlap bias. (Major Point #5)
    • Use of single-SNP instruments and pleiotropy challenges. (Major Point #7, Major Point #9)
  2. Refine overarching conclusions in the Discussion to address discordant protein effects and tissue specificity. (Major Point #4)

    Output

    • Expanded “Study Limitations” subsection.
    • Revised Discussion section.

1st January

Tasks

  1. Ensure all main figures are converted to .tif or .eps format. (Formatting Requirement)
  2. Remove logos or branding from Figures 2, 3, 4, and 6. (Formatting Requirement)
  3. Draft a detailed point-by-point rebuttal letter addressing all reviewer comments. (All Major and Minor Points)

    Output

    • Formatted figures.
    • Draft rebuttal letter.

2nd January

Tasks

  1. Conduct a full review of the manuscript, ensuring consistency and clarity. (All Major and Minor Points)
  2. Check all journal-specific requirements (e.g., formatting, funding statements). (Formatting Requirement)
  3. Finalise the rebuttal letter. (All Major and Minor Points)

    Output

    • Finalised manuscript.
    • Finalised rebuttal letter.

3rd January

Tasks

  1. Generate clean and track-changes versions of the manuscript. (Formatting Requirement)
  2. Prepare supplementary materials and ensure compliance with journal guidelines. (Formatting Requirement)
  3. Submit the revised manuscript and rebuttal letter. (All Major and Minor Points)

    Output

    • Completed submission package.

Reviewer Points

Major Points

  1. Justification for PBMC Inclusion: The manuscript lacks a strong justification for including cellular proteome data from PBMCs in a study focused on CVD, especially given the relevance of plasma proteomics for this context.
  2. Sample Size and Statistical Power: The PBMC cohort (N = 862) is significantly smaller than the UKB plasma proteomics dataset, limiting statistical power. The reliance on single SNP instruments in PBMC analyses further weakens causal inference compared to plasma proteomics, which utilised multiple SNP instruments.
  3. Inconsistent Methodologies: The selection criteria for genetic instruments differ between PBMC and UKB datasets:
    • PBMC analysis used only cis-SNPs without multiple testing correction.
    • UKB data applied Bonferroni correction and multiple SNPs. This inconsistency should either be justified or re-performed with consistent methods.
  4. Opposing Effects of COMT and SWAP70: Proteins like COMT and SWAP70 showed opposing effects in PBMC and plasma data. While methodological explanations are provided, the discussion should explore potential biological reasons, such as tissue-specific protein functions.
  5. Sample Overlap and Bias: Both plasma proteomics and CVD GWAS datasets are derived from UKB, which could introduce biases due to overlapping samples. Sensitivity analyses (e.g., splitting datasets into non-overlapping subsets or using independent datasets for CVD GWAS) are needed to quantify and address potential bias.
  6. Insufficient Information on Genetic Instrument Selection:
    • Missing details on LD ( r^2 ), p-value thresholds, and allele frequency differences used in instrument selection.
    • Lack of clarity about when and how clumping was performed, including thresholds used.
  7. HEIDI Outlier Filtering: The threshold used for identifying and excluding horizontally pleiotropic SNPs in HEIDI filtering was not reported.
  8. Limitations of the Study: The discussion should explicitly address study limitations, including:
    • Potential biases from reliance on affinity-based and MS-based assays.
    • Challenges interpreting protein abundance in the context of genetic variation.
  9. Additional MR Methods: Incorporating alternative MR methods (e.g., IVW, MR-Egger, weighted median, RAPS, MR-PRESSO) could address potential pleiotropy and validate the robustness of results.
  10. Clarity on Terminology: Key terms (e.g., peptide, isoform, protein group, peptide-specific protein, modification-specific peptide) are inconsistently used and not adequately defined. Consistent definitions and possibly a figure illustrating these concepts would help.

Minor Points

  1. Relevance of PBMC Data: Expand the discussion on the relevance of immune cells (PBMCs) to CVD traits and potential unique insights they might offer compared to plasma proteomics.
  2. Contradictory Statistics: Clarify inconsistencies, such as the number of proteins reduced from 127 to 165 after filtering for cis-pQTLs.
  3. Effect Sizes in Table 1: Include effect sizes from GSMR analyses to provide a clearer understanding of the direction and magnitude of effects.
  4. Sample Size Discrepancies: Explain discrepancies in sample sizes (e.g., 861 samples analysed by MS vs. 736 individuals included in GWAS).
  5. Choice of Outcomes: Justify the use of self-reported CVD outcomes in UKB instead of ICD codes.
  6. Excluded SNP Counts: Explain why proteins with 0, 2, or 4 SNPs were excluded from analysis.
  7. Supplementary Data:
    • Provide a supplementary table listing all protein groups in GS, not just significant ones.
    • Include genome reference information for BP coordinates in supplementary tables.
  8. Clumping Process: Clarify when clumping was performed, for which cohorts, and the methodology used.