Round 1

Reviewer 1 Report

This manuscript investigates the frequency and impact of the P1104A/TYK2 genetic variant in a COVID-19 cohort from Southern Brazil, contributing to our understanding of host genetic factors in disease severity within an admixed population. The study is well-structured and addresses a relevant topic. However, there are several limitations, particularly concerning statistical power and data generalizability, that need to be addressed before publication.

While the study provides valuable insights, especially regarding the allelic frequency in a Brazilian cohort, the lack of a statistically significant association between the P1104A/TYK2 variant and COVID-19 severity, attributed to a small number of homozygous individuals, limits the strength of its primary conclusion. Additionally, the geographical limitation of the sample size and potential for incomplete clinical data necessitate further consideration. Addressing these points through revised discussion and potentially expanded analyses would significantly strengthen the manuscript.

Specific Comments:

1.  

   Statistical Power and Homozygous Cases: The authors acknowledge the limitation of the small number of homozygous individuals (n=3) and its impact on statistical power for assessing the association with COVID-19 severity. This is a crucial point, and while acknowledged, it could be discussed in greater detail regarding its implications for definitively ruling out an association. The current finding of no significant association (

    

   p=0.26) means that a true effect might exist but was not detected due to insufficient sample size for this specific genotype.

    

   • Suggestion: Elaborate on the implications of the limited statistical power for the P1104A/TYK2 variant, specifically mentioning that the observed lack of statistical significance does not equate to the absence of an effect. Perhaps suggest a power calculation if feasible for future studies to determine the sample size needed to detect a meaningful effect size for the homozygous genotype.

2.  

   Ancestry Bias and Population Databases: The discussion on the differing C allele frequencies compared to "DNA do Brasil," ABraOM, and gnomAD databases is well-presented and accurately attributes the differences to ancestry patterns and geographical origin. This is a strong point of the manuscript.

      

   • Suggestion: Reiterate the importance of considering ancestry in genetic studies, especially in highly admixed populations, and how the findings highlight the need for more representative local databases.

3.  

   Clinical Data Completeness: The authors mention that "incomplete or unavailable medical records" might have contributed to the lack of significant association between COVID-19 severity and chronic kidney disease, chronic heart disease, or chronic pulmonary disease, which are generally recognized risk factors.

    

   • Suggestion: While acknowledging this limitation, consider discussing any efforts made to mitigate this, or if possible, provide a quantitative estimate of the completeness of these specific comorbidities. This would give readers a clearer picture of the data quality.

4.  

   Generalizability of Findings: The study's geographical limitation to a single Brazilian state (Rio Grande do Sul) is highlighted. This is a valid limitation for generalizing findings to the entire Brazilian population.

    

   • Suggestion: Strengthen the concluding remarks by emphasizing that future multi-centric studies across different Brazilian regions with varying ancestry profiles are crucial to confirm and expand upon these findings.

5. Language and Clarity: The manuscript is generally well-written and clear.

   • Suggestion: A minor proofread for occasional grammatical nuances and flow could further enhance readability. For example, in the abstract, "The P1104A variant in the TYK2 gene has been established as the first common monogenic cause of tuberculosis, with recent studies also suggesting a potential role in COVID-19 severity" could be slightly rephrased for smoother flow, though it is grammatically correct.

      

6. Formatting of Tables: Tables 1, 2, and 3 are clear and provide essential data.

   • Suggestion: Ensure consistent alignment and spacing within tables if not already perfect. (This is a very minor point, often handled in final production.)

Author Response

Reviewer 1

Major comments

This manuscript investigates the frequency and impact of the P1104A/TYK2 genetic variant in a COVID-19 cohort from Southern Brazil, contributing to our understanding of host genetic factors in disease severity within an admixed population. The study is well-structured and addresses a relevant topic. However, there are several limitations, particularly concerning statistical power and data generalizability, that need to be addressed before publication.

Answer: We thank the reviewer for their positive evaluation of our manuscript and for recognizing the relevance of investigating host genetic factors as P1104A/TYK2 variant in the admixed Brazilian population. We acknowledge the concern regarding the statistical power and data generalizability. To address these points we have now included a power calculation for the association analyses in the methods section and expand the discussion as explicit in a point by fashion in the manuscript and in detailed comments below.

 

Detailed comments

While the study provides valuable insights, especially regarding the allelic frequency in a Brazilian cohort, the lack of a statistically significant association between the P1104A/TYK2 variant and COVID-19 severity, attributed to a small number of homozygous individuals, limits the strength of its primary conclusion. Additionally, the geographical limitation of the sample size and potential for incomplete clinical data necessitate further consideration. Addressing these points through revised discussion and potentially expanded analyses would significantly strengthen the manuscript.

 

Specific Comments:

1. Statistical Power and Homozygous Cases: The authors acknowledge the limitation of the small number of homozygous individuals (n=3) and its impact on statistical power for assessing the association with COVID-19 severity. This is a crucial point, and while acknowledged, it could be discussed in greater detail regarding its implications for definitively ruling out an association. The current finding of no significant association (p=0.26) means that a true effect might exist but was not detected due to insufficient sample size for this specific genotype.
• Suggestion: Elaborate on the implications of the limited statistical power for the P1104A/TYK2 variant, specifically mentioning that the observed lack of statistical significance does not equate to the absence of an effect. Perhaps suggest a power calculation if feasible for future studies to determine the sample size needed to detect a meaningful effect size for the homozygous genotype.

Answer: We thank the reviewer for this important suggestion. As recommended, we have expanded the discussion of the implications of the statistical power related to the low number of homozygous individuals in our study. We now explicitly state that the lack of statistical significance does not rule out the possibility of an effect in COVID-19 severity susceptibility. To support this, we conducted a power analysis based on the observed parameters of our cohort (minor allele frequency MAF = 2.85%; n = 1,826; case rate calculated as the number of severe cases (636) divided by the total sample size (1,826) = 34.83%), which revealed that the power to detect an odds ratio of 3.75 under a recessive model, at a significance level of 0.05, was only 12.38%. Additionally, we also performed a sample size calculation, as suggested, which revealed that under the same conditions, approximately 23.000 individuals would be required to achieve 80% power. These results have been incorporated into Methods 2.4 section, line 139. “To assess the statistical power for detecting genetic associations for the P1104A/TYK2 variant and COVID-19 severity, we performed a post hoc power analysis using the genpwr R package (version 1.0.4), based on the observed parameters of our cohort (minor allele frequency = 2.85%; n = 1,826; case rate = 34.83%), under a recessive model, assuming an odds ratio of 3.75 and a significance level of 0.05. Additionally, using the same package, we performed a sample size calculation to estimate the number of individuals required to achieve 80% power, under the same parameters”. We also emphasize these points in the Discussion section line 283, “The power analysis revealed that the study had only 12.38% power to detect an odds ratio of 3.75 at a significant level of 0.05 under a recessive model, given the observed parameters. Therefore, the absence of a statistically significant association between P1104A/TYK2 and COVID-19 severity should not be interpreted as conclusive evidence of no effect, but rather as a reflection of limited statistical power. To estimate the sample size required to adequately assess this association, we performed a sample size calculation, which revealed that approximately 23,000 individuals would be required to achieve 80% power, under the same conditions. It is worth noting that a previous study of TB cases and controls [12] reported an ancestry‑adjusted odds ratio of 5, based on more than 100 000 controls.”.

 

1. Ancestry Bias and Population Databases: The discussion on the differing C allele frequencies compared to "DNA do Brasil," ABraOM, and gnomAD databases is well-presented and accurately attributes the differences to ancestry patterns and geographical origin. This is a strong point of the manuscript.
• Suggestion: Reiterate the importance of considering ancestry in genetic studies, especially in highly admixed populations, and how the findings highlight the need for more representative local databases.

Answer: We thank the reviewer for the positive feedback and for the valuable suggestion. In the revised Discussion section, line 247, we have reinforced the importance of considering ancestry when interpreting genetic association studies, particularly in highly admixed populations such as Brazil's. Our findings highlight that the frequency of the C allele in our COVID-19 cohort differs from the “DNA do Brasil,” ABraOM, and gnomAD databases, reflecting both ancestry composition and regional differences. We have also underscored the need for more representative local and population-specific genomic databases to ensure accurate allele frequency estimation and to avoid potential biases in risk prediction and genetic association studies.

 

1. Clinical Data Completeness: The authors mention that "incomplete or unavailable medical records" might have contributed to the lack of significant association between COVID-19 severity and chronic kidney disease, chronic heart disease, or chronic pulmonary disease, which are generally recognized risk factors.
• Suggestion: While acknowledging this limitation, consider discussing any efforts made to mitigate this, or if possible, provide a quantitative estimate of the completeness of these specific comorbidities. This would give readers a clearer picture of the data quality.

Answer: We thank the reviewer for the relevant comment. We acknowledge that incomplete or unavailable medical records could have impacted the assessment of known risk factors, such as chronic kidney disease, chronic heart disease, and chronic pulmonary disease, which are typically associated with COVID-19 severity. To address this, we have expanded the 2.2 section (Methods, line 100) to include a more detailed description of how comorbidity data were collected, "Information on comorbidities was retrieved through a structured query that extracted all ICD codes ever recorded in each patient’s electronic health record. Additionally, for a subset of 450 individuals enrolled in the “DNA do Brasil” project, comorbidity information was collected mainly through manual review of health records, with a small portion also completing a questionnaire administered at the time of informed consent". We have also included the completeness of comorbidity data results Discussion section, line 277, 31.00% for chronic kidney disease (566/1,826), 30.07% for chronic heart disease (549/1,826), and 24.70% for chronic pulmonary disease (451/1,826) to address this limitation and its potential impact on the interpretation of results.

 

1. Generalizability of Findings: The study's geographical limitation to a single Brazilian state (Rio Grande do Sul) is highlighted. This is a valid limitation for generalizing findings to the entire Brazilian population.
• Suggestion: Strengthen the concluding remarks by emphasizing that future multi-centric studies across different Brazilian regions with varying ancestry profiles are crucial to confirm and expand upon these findings.

Answer: We thank the reviewer for this important comment. We recognize that the geographical limitation of our study to a single Brazilian state may restrict the generalizability of our findings to the broader Brazilian population, as Brazil is a country with continental dimensions and significant regional variation in genetic ancestry, healthcare access, and socioeconomic conditions. In response to this suggestion, we have revised the concluding remarks, line 304, to emphasize the need for future multi-centric studies encompassing different Brazilian regions with diverse ancestry profiles, which will be crucial to confirm and expand upon our findings.

 

1. Language and Clarity: The manuscript is generally well-written and clear.
• Suggestion: A minor proofread for occasional grammatical nuances and flow could further enhance readability. For example, in the abstract, "The P1104A variant in the TYK2 gene has been established as the first common monogenic cause of tuberculosis, with recent studies also suggesting a potential role in COVID-19 severity" could be slightly rephrased for smoother flow, though it is grammatically correct.

Answer: We thank the reviewer for the positive feedback regarding the clarity of the manuscript. As suggested, we have performed an additional round of proofreading to address minor grammatical nuances and improve the overall flow of the text.

 

1. Formatting of Tables: Tables 1, 2, and 3 are clear and provide essential data.
• Suggestion: Ensure consistent alignment and spacing within tables if not already perfect. (This is a very minor point, often handled in final production.)

Answer: We thank the reviewer for pointing this out. We carefully reviewed the Tables to ensure consistent formatting.

Author Response File: Author Response.docx

Reviewer 2 Report

This study does not have a sufficient sample size to perform the statistical tests needed to determine whether P1104A/TYK2, rs34536443 is a risk factor for severe COVID-19 in this population of patients. While the question and hypothesis is reasonable, the study does not contribute to a greater understanding of how this variant is contributing to increased disease risk or increased severity of COVID-19 infection. 

I have no detailed comments.

Author Response

Reviewer 2

 

Does the introduction provide a comprehensive yet concise overview about the state of knowledge in the area of research?

No

P1104A/TYK2, rs34536443 is a polymorphism, that reduces TYK2 signalling compared to the alternative allele. However, while it increases susceptibility to TB it should be seen in the context as a risk factor rather than its cause. The wording needs to be amended to make it clear that carriers of P1104A have a reduced signalling capacity, which is disadvantageous in the context of some infections, namely TB, but is highly protective against several autoimmune diseases.

Answer: Thank you for your comment, we have revised the text to clarify that P1104A (rs34536443) is a polymorphism that reduces TYK2 signaling compared to the alternative allele. We have also rephrased the statement about tuberculosis to indicate that P1104A acts as a mendelian susceptibility factor for TB in homozygotes, with incomplete penetrance. We also acknowledge its protective role against several autoimmune disorders (Introduction, line 73). These changes provide a more accurate and nuanced description of the variant’s functional and clinical relevance.

 

Are the results presented clearly and in sufficient detail, are the conclusions supported by the results and are they put into context within the existing literature?

No

The sample size is too small to make any conclusions. The severe and non-severe cases are not comparable - age, sex, comorbidities - to allow an analysis to be performed that fairly measures genetic variation.

Answer: We thank the reviewer for this important observation. We acknowledge that the small number of homozygous individuals and the differences in age, sex, and comorbidities between severe and non-severe groups limit the robustness of our conclusions. To address this, we have expanded the Results and Discussion sections to more clearly highlight these limitations and emphasize that our findings should be interpreted as exploratory rather than conclusive. We also included a post hoc power analysis demonstrating that our study has limited statistical power (12.38% for OR = 3.75 under a recessive model) and that approximately 23,000 individuals would be required to achieve 80% power under the same assumptions. Additionally, we have contextualized our results within the existing literature, stressing the novelty and relevance of examining the P1104A/TYK2 variant in an admixed population like Brazil, where such data are scarce.

 

Does this article provide a relevant contribution to the scientific discussion of this topic?

No

There are no conclusive findings.

Answer: We acknowledge that we have limited statistical power to draw definitive conclusions regarding the role of P1104A/TYK2 as a risk factor for severe COVID-19, as explicit above. However, our study provides the first reports of this variant in a highly admixed Brazilian population, which remains underrepresented in global COVID-19 genomic studies. Large GWAS consortia, including the COVID-19 Host Genetics Initiative, have previously highlighted TYK2 variants, including P1104A, as potential contributors to severe COVID-19. Our results complement these findings by reporting allele frequencies and genotypic distributions in a genetically diverse cohort, reinforcing the need for larger studies in admixed populations to better understand the role of this variant in COVID-19 susceptibility and severity.

 

Major comments

This study does not have a sufficient sample size to perform the statistical tests needed to determine whether P1104A/TYK2, rs34536443 is a risk factor for severe COVID-19 in this population of patients. While the question and hypothesis is reasonable, the study does not contribute to a greater understanding of how this variant is contributing to increased disease risk or increased severity of COVID-19 infection. 

Answer: We thank the reviewer for this comment. We agree that our sample size, particularly the small number of homozygous carriers, limits the statistical power to establish a definitive association between P1104A/TYK2 and COVID-19 severity. As addressed in the revised manuscript, we now include a post hoc power analysis showing that the power to detect an OR of 3.75 under a recessive model was only 12.38%, and that a sample size of approximately 23,000 would be required to reach 80% power. These limitations are explicitly discussed, and the findings are presented as exploratory. Given the scarcity of genetic data from admixed Latin American populations, our study contributes novel insights that can inform and guide future large-scale studies.

Author Response File: Author Response.docx

Reviewer 3 Report

The manuscript addresses the relevance of the TYK2 P1104A variant in COVID-19 severity in a large Brazilian cohort. The topic is relevant and contributes to the understanding of host genetic susceptibility in admixed populations, which are typically underrepresented in genomic studies. The methods are appropriate, and the statistical analyses are clearly reported.

However, some clarifications and improvements are needed before the manuscript can be considered for publication:

• Given the rarity of the CC genotype, the authors should provide a power analysis to support the claim that the sample size is insufficient to detect associations. This is especially important to contextualize the lack of statistically significant results.
• The observed higher prevalence of obesity, diabetes, and hypertension in severe cases aligns with previous reports. However, it would be useful to include multivariate analyses adjusting for these confounders when evaluating the association between genotype and disease severity.
• The authors appropriately reported the lack of significant association between P1104A and disease severity. However, it would benefit from a deeper discussion on possible immunological reasons for this negative result, particularly in relation to the distinct immunological effects of homozygous P1104A versus complete TYK2 deficiency.
• The authors mention excluding overlapping samples from “DNA do Brasil” but should clearly specify the number of excluded individuals and assess whether this impacts frequency comparisons.
• “Non-severe” is a broad category; consider whether it is possible to subcategorize this group (e.g., mild vs. moderate) if data allow.
• Several minor typographical issues are present (e.g., inconsistent spacing, line breaks). These should be corrected in the final version.

Author Response

Reviewer 3

 

Major comments

The manuscript addresses the relevance of the TYK2 P1104A variant in COVID-19 severity in a large Brazilian cohort. The topic is relevant and contributes to the understanding of host genetic susceptibility in admixed populations, which are typically underrepresented in genomic studies. The methods are appropriate, and the statistical analyses are clearly reported.

Answer: We thank the reviewer for recognizing the relevance of our study. Investigations about host genetic susceptibility in admixed populations, such as the Brazilian cohort, is crucial to filling gaps in global genomic studies. We have ensured that our methods and analyses are clearly described and have further refined the discussion to emphasize the importance of these findings within the context of underrepresented populations.

 

Detailed comments

However, some clarifications and improvements are needed before the manuscript can be considered for publication:

• Given the rarity of the CC genotype, the authors should provide a power analysis to support the claim that the sample size is insufficient to detect associations. This is especially important to contextualize the lack of statistically significant results.

Answer: We thank the reviewer for this valuable suggestion. To address the concern regarding the rarity of the CC genotype and its implications for statistical power, we performed a post hoc power analysis. This analysis revealed that the study had only 12.38% power to detect an odds ratio of 3.75 under a recessive model. Additionally, a sample size calculation showed that approximately 23,000 individuals would be required to achieve 80% power under the same conditions. These results support the claim that the sample size was insufficient to detect a significant association and reinforce that the absence of statistical significance should not be interpreted as evidence of no effect but rather as limitation in statistical power due to limited sample size. These analyses have been incorporated into the revised manuscript in the Methods section 2.4, line 139, and in the Discussion, line 283.

 

• The observed higher prevalence of obesity, diabetes, and hypertension in severe cases aligns with previous reports. However, it would be useful to include multivariate analyses adjusting for these confounders when evaluating the association between genotype and disease severity.

Answer: We appreciate the reviewer’s suggestion regarding the inclusion of a multivariate analysis adjusting for potential confounders such as obesity, diabetes, and hypertension. We fully agree that these variables are important to consider when evaluating this association, to better isolate the effect of the genotype, independent of underlying clinical risk. However, due to limited completeness of comorbidity data in our cohort, adjusting for these variables in a multivariate model would substantially reduce the sample size, retaining only approximately 25% of the original sample (450 out of 1,826 individuals). Instead, we performed a multivariate analysis adjusting for age and sex, which are also associated with COVID-19 severity and for which we had complete data for the entire cohort. This approach allowed us to control for major demographic confounders without compromising sample size. This has now been included in the revised manuscript 2.4 section (Methods, line 136), “Logistic regression models were fitted to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the association between the P1104A/TYK2 genotype and COVID-19 severity, adjusting for age and sex, known risk factors for severe COVID-19”. The corresponding results are presented in the Results section, line 183, and in Table 4 (line 188).

 

• The authors appropriately reported the lack of significant association between P1104A and disease severity. However, it would benefit from a deeper discussion on possible immunological reasons for this negative result, particularly in relation to the distinct immunological effects of homozygous P1104A versus complete TYK2 deficiency.

Answer: Thank you for this valuable suggestion. We have expanded the discussion to better address the potential immunological reasons underlying the lack of association between the P1104A variant and COVID-19 severity, including a more detailed comparison between the immunological effects of homozygous P1104A and complete TYK2 deficiency. The revised text can be found in the Discussion section, line 204: “While previous studies have shown that this variant may impair IL-12 and IFN-I signaling in some experimental settings [17], the physiological relevance appears limited. Indeed, individuals with complete TYK2 deficiency typically exhibit impaired responses to IL-12, IL-23, and IFN-I, resulting in increased susceptibility to intracellular bacteria and a wide range of viral infections, including herpesviruses [6]. In contrast, homozygous carriers of the P1104A variant exhibit impaired IL-23 responses, while IFN-I signaling appears only partially affected [10,12]. This partial functional impairment may be sufficient to predispose to mycobacterial infections, such as tuberculosis, but insufficient to significantly compromise antiviral immunity, particularly SARS-CoV-2, which relies heavily on early and effective IFN-I responses [16,17]. This might also explain the absence of a strong statistically significant association between P1104A variant and COVID-19 severity”.

 

• The authors mention excluding overlapping samples from “DNA do Brasil” but should clearly specify the number of excluded individuals and assess whether this impacts frequency comparisons.

Answer: We thank the reviewer for this important comment. To address it, we have now explicitly stated in section 2.4 section (Methods, line 131) that 485 overlapping samples were excluded from the “DNA do Brasil” database to avoid data duplication. We also strengthen the Discussion section, line 242, informing that the frequency of the C allele in the “DNA do Brasil” database decreased from 2.00% to 1.81% after removal of these samples, likely reflecting the higher proportion of European ancestry among individuals from Rio Grande do Sul within our COVID-19 cohort compared to the overall “DNA do Brasil” dataset.

 

• “Non-severe” is a broad category; consider whether it is possible to subcategorize this group (e.g., mild vs. moderate) if data allow.

Answer: We appreciate the reviewer’s suggestion. However, further subcategorization of the non-severe group into mild or moderate cases is not feasible with the available data. In our cohort, clinical severity was assessed based on whether the patient required ICU admission, ventilation, progressed to acute respiratory distress syndrome (ARDS), and the final outcome (discharge or death). Unfortunately, we do not have access to additional clinical parameters, such as oxygen saturation measures, oxygen flow rate, or other detailed indicators of disease severity that would allow for more refined classification.

 

• Several minor typographical issues are present (e.g., inconsistent spacing, line breaks). These should be corrected in the final

Answer: We thank the reviewer for pointing this out. We carefully reviewed the entire manuscript and corrected the typographical issues.

 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

This is a much more measured interpretation and explanation of the results. While it does not have sufficient power to really test the influence of this variant to its fullest, the data provides an important step towards understanding how its function may be exerted across different ancestral populations, which is vital when predicting and managing the the potential impact and severity of infections such as COVID-19.

none

Reviewer 3 Report

The authors have responded to all comments

The authors have responded to all comments