BDP1 Expression Correlates with Clinical Outcomes in Activated B-Cell Diffuse Large B-Cell Lymphoma
Round 1
Reviewer 1 Report
In the article titled "BDP1 Expression Correlates with Clinical Outcomes in Activated B-cell Diffuse Large B-Cell Lymphoma", authors describe the gene expression analysis of BDP1 from multiple patients from different sources, its possible role in Non-Hodgkin Lymphoma (NHL), more specifically in a subtype of the lymphoma, activated B-cell (ABC) diffuse large B-cell lymphoma (DLBCL). In a significant contribution to the much needed lymphoma research, the paper addresses the deregulation of TFIIIB subunit IBP1 and its role in the predictability of patient outcome.
Here are few points to consider:
1. The paper uses an in-depth gene expression analysis while the more standardized, meta-analysis Effectsize and FDR could also be evaluated for the mentioned genes.
2. BDP1 is significantly overexpressed in colorectal cancer and breast cancer while under expressed significantly in the lymphoma studied in this paper.
The context for overexpression of BDP1 needs to be explained while contrasting for under expression in lymphoma.
3. A minor point, in general the figures needs bit more clarity for axis labels, for example Fig 2 A and B.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
In this manuscript, Cabarcas-Petroski et al. used the processed gene expression data from oncomine to show the decreased expression of the BDP1 gene in Non-Hodgkin Lymphomas and its correlation with various clinical features. Overall the manuscript is well written, but the authors need to carry out more analysis to verify their results. Below are my comments and suggestions to improve the current analysis.
Comments to Authors:
- The oncomine research edition website is going offline on 17th January 2022. How will other researchers reproduce their results in the future?
- Authors should briefly write about the data processing/analysis steps (including normalization, batch effect correction, differential expression analysis, etc.) used by them or in the oncomine database.
- The authors should also carry out pathways/gene ontology enrichment analysis to show the significant enrichment of the genes in a particular pathway instead of picking a few genes in their analysis.
- The study seems to be biased on the dataset in the oncomine database. Have the authors tried to verify their results based on other datasets (RNASeq or microarrays)?
- Did the authors carry out the multiple hypothesis testing correction (FDR, BH, etc) after carrying out the differential expression analysis? Is the p-value reported is the raw one or the corrected one?
- The authors should also show the correlation between the genes (and significance p-value) to verify their claim about significant correlation. Currently, only the differential expression analysis’s p-values are reported.
- In the results section 3.5, the authors only showed the results from Eckerle Lymphoma datasets, not the Shaknovich Lymphoma dataset. They should show results from both datasets.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
RNA polymerase III plays indispensable role in regulating eukaryotic cell proliferation. While the deregulation of human TFIIIB subunits, BRF1 and BRF2 are well reported in human cancers, it is unclear that if the BDP1 subunit, which is shared by the two forms of human TFIIIB that have been identified, is universally deregulated in human cancers. In this work, by performing a meta-analysis of patient data in the Oncomine database to analyze BDP1 alterations in human cancers, the authors found BDP1 is significantly altered in a subset of human cancers, then further revealed that decreased BDP1 expression may correct with clinical outcomes in activated B-cell diffuse large B-cell lymphoma, and proposed BDP1 could be a novel target for therapeutic intervention for patients with NHL subtypes. This work is clearly developed and presented, rendering further clinical investigation of the role of BDP1 in human cancers. I have no major comments, suggestions from my side are that, the “dataset description” (line 109) may need to be presented as a Table, and it would be better if the “Materials and Methods” part could be described more adequately.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
I would like to thank the authors for answering my questions. I have additional comments and suggestions below.
Comments to Authors:
- On a quick search, I was able to find another microarray dataset of ALCL cells (https://www.ebi.ac.uk/arrayexpress/experiments/E-TABM-117/). The authors can use this data (and search for other datasets apart from oncomine) to verify that results can be generalized and is not dependent on data from one lab.
- The authors have incorrectly marked two sections as 3.5. In my previous comments, I referred to the results section “Decreased BDP1 expression correlates with FOXP1 and BCL6 expression”. The authors only showed the results from Eckerle Lymphoma dataset (ALK+ ALCL data), not the Brune dataset (BL data). They should show results from both datasets. Also, in figure 4, are the rank, p-values, and fold-change come from differential expression analysis? If yes, then the authors should do a correlation test (Pearson or Wilcoxon) to show that the expressions of the genes are correlated, which is stated in the heading of this section.
- What is the "median gene rank" used in the analysis? Is the gene rank based on the normalized gene expression values?
- The authors should mention the type of correlation used in Figure 6.
- Please increase the font size of the text in the figures. For example, the names of the genes in the heatmaps or the axis titles in the boxplot are not clearly visible. The authors should also improve the quality of the figures as the text is breaking when I tried to zoom them.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 3
Reviewer 2 Report
Based on the author's response it seems like oncomine is the only database that has curated microarray database from public repositories. I want to point out that the oncomine papers that the authors are referring to are from the years 2004 and 2007. I believe the oncomine database has been updated regularly but the core data processing and data curation remain the same (and if they have updated their curation method then the authors should write or cite that). It means that there must ways to improve the data curation step. This was verified when I tried to search for datasets in other curated databases and found one. On a quick search, I was able to find a plethora of database that has curated datasets from publicly available databases (eg. https://academic.oup.com/database/article/doi/10.1093/database/baab006/6143045). The authors can themselves search for them and should not assume that I want them to use a specific one. Instead of acknowledging that, the authors seem to find faults in the data submission. We all know that the metadata of studies are not 100% accurate and that is the reason there are various strategies employed to curate microarray/sequencing public datasets (including machine learning, manual curation). My major point was to cross-check the results that the authors are making in sections 3.5 and 3.6. It seems like the authors are quite inflexible to cross-check the publicly available datasets in other latest curated databases other than oncomine. Also, the authors in their comments reference comments from Reviewers 1 and 2. The authors should understand the whole reason for selecting more than one reviewer is to get comments from researchers from different domains. It is an integral part of the review process. I tried to help the authors to improve their paper, but it seems like they are not flexible.
Author Response
I submitted a reply to the editor as requested addressing the comments from reviewer 2, round 3, and the editors comments.
