Two-step light gradient boosted model to identify human west nile virus infection risk factor in Chicago

Guangya Wan; Joshua Allen; Weihao Ge; Shubham Rawlani; John Uelmen; Liudmila Sergeevna Mainzer; Rebecca Lee Smith

doi:10.1371/journal.pone.0296283

Peer Review History

Original SubmissionJune 8, 2023
18 Sep 2023 Decision Letter - José Ramos-Castañeda, Editor PONE-D-23-17817Two-Step Light Gradient Boosted Model to identify human West Nile Virus infection risk factor in ChicagoPLOS ONE Dear Dr. Ge, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. Please submit your revised manuscript by Nov 02 2023 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. Please include the following items when submitting your revised manuscript: A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'. An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'. If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols. We look forward to receiving your revised manuscript. Kind regards, José Ramos-Castañeda, M.Sc., Ph.D Academic Editor PLOS ONE Journal Requirements: When submitting your revision, we need you to address these additional requirements. 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf 2. We note that the grant information you provided in the ‘Funding Information’ and ‘Financial Disclosure’ sections do not match. When you resubmit, please ensure that you provide the correct grant numbers for the awards you received for your study in the ‘Funding Information’ section. 3. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability. Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized. Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access. We will update your Data Availability statement to reflect the information you provide in your cover letter. 4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information. Additional Editor Comments: I have read the manuscript in detail and consider that this manuscript should be extensively reviewed and revised. One of the reviewers has mentioned in detail aspects that should be considered in the revised manuscript. [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes Reviewer #2: Partly ******** 2. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes Reviewer #2: Yes ****** 3. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes Reviewer #2: Yes ****** 4. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: Yes Reviewer #2: Yes ****** 5. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: The author is requested to include quantitative values within the conclusions. Thinking of concluding based on the initial objectives of the project. On the other hand, it is requested that the graphics be inserted in a more readable format, the images are pixelated. Please improve the quality of the figures. Within the text figure 6 remains to be detailed. Regarding the utility of the program as a predictive model, it is a very valuable contribution. Reviewer #2: Wan et al. performed a two-step modeling approach to predict WNV in the Chicago area and identify important parameters for prediction. Work to improve WNV prediction is vital in protecting human health so this work could be a notable addition. Their approach appears to build off a previous model by Karki et al., but the motivation for this is not well defined nor is the improvement obtained by this model over the previous results. Also, the methods and results are often difficult to follow and only a cursory discussion is made of the results. The data selection and model building procedures could benefit from clarification (e.g., rewording, visual diagram of the steps in the process). I have several comments and questions that need to be addressed to improve the understanding and potential utility of the results for WNV prediction. Line 29-59. Please consider re-structing or breaking up this single paragraph into smaller paragraphs to aid the reader in following the multiple topics included. As I see it, this paragraph mentions previous predictive models as well as the relationship of climate, land cover, socioeconomic factors with WNV infection in mosquitoes and humans in a list-like fashion (i.e., this author found this list of covariates as significant and this other author found this list of covariates significant). Previously developed predictive models are then re-visited in more depth in the following paragraph. Line 35. The statement is incorrect. Hahn et al. found a nonsignificant relationship between above-average temperatures and WNV risk in the western regions. Line 48. The term “geologic” might be used incorrectly when referring to land cover and types of land use. Line 51. Please provide more explanation on how ethnicity is a key predictor of WNV circulation in humans as there is no direct causal relationship between ethnicity (or race) and disease incidence. You indicate in Line 59 that socioeconomic differences between racial groups explained differences in prevalence. Line 68. Please provide more information in light GBM models. The current description would describe random forests as well, but the following sentence indicates that there are differences in fitting speed. How are light GBM and random forest models different? Line 75. Do you mean socioeconomics instead of social economics? Line 83. Assuming the data (WNV case, mosquito surveillance) are available for the subsequent years, could you use your model trained on 2005-216 to predict more recent years as an external validation of model performance? Line 93. Why was weather data aggregated from the native grid to the census tract and then to the hexagonal grid? Why not aggregate directly from the original grid to the hexagonal grid without the potential loss in information across multiple aggregation steps of different spatial scales? Line 109 (Table 1). Were all MIR variables the mean during the time period (0-4 weeks) or just at the 0-week lag? Similarly, were the temperature variables mean, min, or max during the time periods? Also, the “nature” category seems a bit odd to me. It seems more appropriate/understandable to break it into two categories: mosquito surveillance and weather. Line 114. This statement is confusing. Do you mean that the KS test is not designed to test for collinearity among predictors? Line 122. The statement “therefore, it is possible that the factors selected in the model are correlated with the true predictors” is confusing. What do you mean by “true predictors”? Line 124. What metric did you use to identify functional significance of covariates for inclusion in the model? Line 128. Please provide more details on the model building and validation. The Results (Line 190) indicate that cross-validation was performed, but no details are given here in the Methods. Similarly, what hyperparameters were selected from the grid search? Line 155. The KS test is not designed to assess causality. Please rephrase your statement to something like the following “characteristics were not significantly related to WNV risk”. Line 159, 195. The mention of the four main + one residual categories in the figure legends is confusing as only three categories have been mentioned in the text and Table 1. Line 181. Please list what the variables with the highest KS scores were. These variables could be gleaned from the figures, but it is unclear in the text which of the medium to highly correlated variables were included in the model. Line 191. Please indicate which metric was used to discern variable importance as multiple metrics could have been used. This is stated in the legend for Fig 3, but not in the text itself. Line 206. Consider re-wording to “We observed a natural break in variable importance (approx.. 20) between MIR at 3-week lags (mirlag3) and percentage of open water (owpct)” to clarify the meaning and align with the statement in Line 222. Lines 209 and 230. Clarify the feature selection resulting from maximizing TPR and FPR in the test set. It is not clear if any features were removed during this step. Line 215. It appears that the results in Table 2 are from the full model (Line 210) as well as from the reduced model (Line 223). Please clarify which model was used and provide further results for the other model’s performance on the test dataset. Line 229. Given variation in the runs, it could be prudent to provide some indication in the plots (Figs 3-4) on the magnitude of variation (e.g., 95% CI or ranges). Line 247. Were partial dependence plots produced for all covariates in the full or reduced model? Line 249 (Figs 5-7). Does the shaded region represent a 95% CI of the marginal effects? If so, what is the significance of this interval containing 0 for most of the covariates? Also, please provide units for covariates (x-axis) and marginal effect (y-axis). Line 259. Please provide support for your supposition that warmer January temperatures increase mosquito survival over the winter. Line 281. This conclusion appears to contradict your statement in Line 243 that the reduced model was sufficient to describe the results. Line 285. To what are you comparing the recall of the light GBM model? Additionally, how does this statement relate to the discussion of importance features that follows? Line 288. Please expound on what these consistent relationships are instead of just citing references and leaving it up to the reader to deduce what these relationships are. Line 290. Why might income have been identified as an important variable and yet have no marginal effect? Line 295. Given the effort taken to reduce collinearity among predictors, how much correlation remained between lagged MIR and lagged temperature that could mask the relationship of socioeconomic factors and WNV incidence? Line 309. Given the outlined conclusions, I do not see how the results of this model enhance the ability to predict WNV incidence. Please provide more details on how this study could develop or modify currently existing thresholds for public health. ****** 6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No Reviewer #2: No ******** [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. https://doi.org/10.1371/journal.pone.0296283.r001
Revision 1
18 Nov 2023 Author Response Reviewer #1: The author is requested to include quantitative values ??within the conclusions. Thinking of concluding based on the initial objectives of the project. On the other hand, it is requested that the graphics be inserted in a more readable format, the images are pixelated. Please improve the quality of the figures. Within the text figure 6 remains to be detailed. Regarding the utility of the program as a predictive model, it is a very valuable contribution. Reviewer #2: Wan et al. performed a two-step modeling approach to predict WNV in the Chicago area and identify important parameters for prediction. Work to improve WNV prediction is vital in protecting human health so this work could be a notable addition. Their approach appears to build off a previous model by Karki et al., but the motivation for this is not well defined nor is the improvement obtained by this model over the previous results. Also, the methods and results are often difficult to follow and only a cursory discussion is made of the results. The data selection and model building procedures could benefit from clarification (e.g., rewording, visual diagram of the steps in the process). I have several comments and questions that need to be addressed to improve the understanding and potential utility of the results for WNV prediction. Line 29-59. Please consider re-structing or breaking up this single paragraph into smaller paragraphs to aid the reader in following the multiple topics included. As I see it, this paragraph mentions previous predictive models as well as the relationship of climate, land cover, socioeconomic factors with WNV infection in mosquitoes and humans in a list-like fashion (i.e., this author found this list of covariates as significant and this other author found this list of covariates significant). Previously developed predictive models are then re-visited in more depth in the following paragraph. Sure. We are breaking up the paragraph and modifying the texts. In this paragraph, our goal is to list factors influencing human incidences from the literature. In the literature we found the weather, land cover, and socioeconomic factors. The impacts of these factors on infection in mosquitoes are also considered. The models, mainly linear regression and tree methods, are discussed in the next paragraph. Line 35. The statement is incorrect. Hahn et al. found a nonsignificant relationship between above-average temperatures and WNV risk in the western regions. Thank you. We have double-checked the paper by Hahn et. al. and modified the statement. Line 48. The term “geologic” might be used incorrectly when referring to land cover and types of land use. Thank you. The term is corrected to “land cover”. Line 51. Please provide more explanation on how ethnicity is a key predictor of WNV circulation in humans as there is no direct causal relationship between ethnicity (or race) and disease incidence. You indicate in Line 59 that socioeconomic differences between racial groups explained differences in prevalence. According to Hernandez et al., the racial disparity of the disease incidence might be due to underreport affected by health insurance and willingness to seek medical care. Moreover, there are behavioral risk associated with different ethnic groups, like whether the individuals working outside home. Line 68. Please provide more information in light GBM models. The current description would describe random forests as well, but the following sentence indicates that there are differences in fitting speed. How are light GBM and random forest models different? Yes, we agree that our previous description is not accurate for the lightGBM model, which is not an ensemble tree method. The lightGBM model is similar to random forest in that both models select features based on decision trees. However, LightGBM, as its name indicated, is a gradient boosting algorithm, that the decision trees are trained based on the errors of the previous trees. On the other hand, random forest is an ensemble method tha trains multiple decision trees independently, and then combines the predictions of the trees via averaging or voting. The random forest used a depth-wise strategy for growing trees while lightGBM used a leaf-wise strategy. Therefore, the trees by random forest are more balanced, and the lightGBM approach converges faster. Line 75. Do you mean socioeconomics instead of social economics? Yes, we corrected “social economics” to “socioeconomics” Line 83. Assuming the data (WNV case, mosquito surveillance) are available for the subsequent years, could you use your model trained on 2005-216 to predict more recent years as an external validation of model performance? Yes, our model can also be used to train and predict more recent years. The WNV case number can be obtained from local health department. However, the weekly mosquito infection rate at the 1km hexagons needs to be collected. Line 93. Why was weather data aggregated from the native grid to the census tract and then to the hexagonal grid? Why not aggregate directly from the original grid to the hexagonal grid without the potential loss in information across multiple aggregation steps of different spatial scales? Yes, we checked back with Karki, et.al. and found we have made a mistake. The weather data are aggregated to the hexagonal grids directly from the 4km resolution PRISM data. The socioeconomics data are downloaded at the census tract level from the census beaureau. Line 109 (Table 1). Were all MIR variables the mean during the time period (0-4 weeks) or just at the 0-week lag? Similarly, were the temperature variables mean, min, or max during the time periods? Also, the “nature” category seems a bit odd to me. It seems more appropriate/understandable to break it into two categories: mosquito surveillance and weather. We have modified Table 1, and make the general categories same as the Table 1 in Karki. et.al., that is: Land cover, Mosquito infection rate, Weather, and Demographic factors. MIRmean, MIRlag1-4, preci, precilag1-4, tempc, templag1-4 are weekly means. Line 114. This statement is confusing. Do you mean that the KS test is not designed to test for collinearity among predictors? Yes, KS test is not to test the collinearity and correlation between the predictors, therefore, we calculated the Pearson’s correlation and filtered out more features before we started modeling. Line 122. The statement “therefore, it is possible that the factors selected in the model are correlated with the true predictors” is confusing. What do you mean by “true predictors”? By “true predictors”, we wanted to say that the strongest features identified by the model are not necessarily the causal factors for the WNV. Line 124. What metric did you use to identify functional significance of covariates for inclusion in the model? We used Pearson’s correlation to group the covariates. Then, in each group, we selected the ones with the highest -log(p-value) in the KS test. Line 128. Please provide more details on the model building and validation. The Results (Line 190) indicate that cross-validation was performed, but no details are given here in the Methods. Similarly, what hyperparameters were selected from the grid search? We used a randomized search with a 5-fold cross validation. We added Table 2 for the hyperparameter range we are searching. Line 155. The KS test is not designed to assess causality. Please rephrase your statement to something like the following “characteristics were not significantly related to WNV risk”. Thank you. We rephrased the statements. Line 159, 195. The mention of the four main + one residual categories in the figure legends is confusing as only three categories have been mentioned in the text and Table 1. The figures are regenerated to match the categories to the ones in Table 1. Line 181. Please list what the variables with the highest KS scores were. These variables could be gleaned from the figures, but it is unclear in the text which of the medium to highly correlated variables were included in the model. We include KS results in supplementary material 4 now. Line 191. Please indicate which metric was used to discern variable importance as multiple metrics could have been used. This is stated in the legend for Fig 3, but not in the text itself. We used the Gini feature importance to discern variable importance. Line 206. Consider re-wording to “We observed a natural break in variable importance (approx.. 20) between MIR at 3-week lags (mirlag3) and percentage of open water (owpct)” to clarify the meaning and align with the statement in Line 222. Thank you for the suggestion. We re-run the lightGBM analysis for 5 times, took the 6 data sets, calculated the error values and means of each feature. Then we ranked the features by feature importance from high to low. We performed t-test between each neighboring features and found at rank 16, the features are significantly different from each other (p=0.03) Lines 209 and 230. Clarify the feature selection resulting from maximizing TPR and FPR in the test set. It is not clear if any features were removed during this step. Thank you for pointing this out. We wrongly stated the threshold is for feature selection. But the threshold is to determine whether a WNV case occurs in a hexagon grid in a week. No features were removed during this step. With the models, we predict the probabilities of a WNV case occurrence in a 1km hexagon grid during a week. If the probability is greater than equal to the threshold, we define the WNV case occurs. If the probability is less than the threshold, we define the WNV case doesn’t occur. Then the program checks the prediction with the data set to see if they match, and categorized the result in to true positive, true negative, false positive, and false negative. Then the difference of TPR and FPR are calculated for each threshold. The threshold for the largest difference of TPR and FPR are selected to determine if a hexagon grid has WNV case occurrence. Line 215. It appears that the results in Table 2 are from the full model (Line 210) as well as from the reduced model (Line 223). Please clarify which model was used and provide further results for the other model’s performance on the test dataset. Thank you. We have corrected the tables and the text. The “table 2” (new table 3) is the confusion matrix for the full model at the best threshold for predicted probability, and the “table 3” (new table 4) is the reduced model at the best threshold for predicted probability. Line 229. Given variation in the runs, it could be prudent to provide some indication in the plots (Figs 3-4) on the magnitude of variation (e.g., 95% CI or ranges). We have added the 95% CI lines for figure 3. However, the reduced model does not have any repetitive runs. Therefore figure 4 doesn’t have 95% CI lines. Line 247. Were partial dependence plots produced for all covariates in the full or reduced model? The partial dependence plots are produced for the full model Line 249 (Figs 5-7). Does the shaded region represent a 95% CI of the marginal effects? If so, what is the significance of this interval containing 0 for most of the covariates? Also, please provide units for covariates (x-axis) and marginal effect (y-axis). Thank you for pointing this out. Yes, we have made a mistake in generating the partial dependence plot. The shaded region represents the lines of all predicted outcome lines. We have changed the settings and produced the figures again. The yellow shade around the center line is the variability of the central tendency. The extra lines are the predicted outcome lines. Since the data is mainly WNV_binary == 0, therefore there will always been predicted outcome lines lying on the y==0 place, if the data sampled to produce the lines all have WNV_binary==0. Therefore, if we plot the full range, it will always pass 0s. In the new figures, the central yellow shade shows the central tendency, and the trend is clearer. Line 259. Please provide support for your supposition that warmer January temperatures increase mosquito survival over the winter. Hahn, et.al. has found that warmer December and January temperatures correlated with the WNV incidences in the later months. They suggested that a warmer winter will increase mosquito and egg survival in the winter. Poh, et.al. has also found warmer winter has associated with mosquito abundance as well as WNV outbreak 8 months later. Line 281. This conclusion appears to contradict your statement in Line 243 that the reduced model was sufficient to describe the results. Yes, we agree. We have changed the analysis to the full models. Line 285. To what are you comparing the recall of the light GBM model? Additionally, how does this statement relate to the discussion of importance features that follows? Yes, we agree with the comment. The linear models do not provide a recall. Therefore, we cannot compare the recall of the lightGBM models to the linear models. The sentence is removed and the whole paragraph now focuses on important feature discussion Line 288. Please expound on what these consistent relationships are instead of just citing references and leaving it up to the reader to deduce what these relationships are. Yes. We have added these in the discussion.; Line 290. Why might income have been identified as an important variable and yet have no marginal effect? Income was selected as one of the top features in our model. However, according to the new partial dependency plot, its marginal effect is very low. Its effect first decreases and then increases. We believe income disparity of WNV incidences is similar to what Hernandez et al. described for racial disparity. In the group lower than the median $65,000, it is likely that people are reluctant to seek medical care when WNV symptoms are mild, therefore having an under reporting. The slight positive effect following it might be related to working location and other behavioral risks. Line 295. Given the effort taken to reduce collinearity among predictors, how much correlation remained between lagged MIR and lagged temperature that could mask the relationship of socioeconomic factors and WNV incidence? Our statement that the socioeconomic and land cover effect was masked was wrong. We aggregated the data by year and calculated the models again, and did not find any strong effect. We have to say that the socioeconomic and land cover really have weaker effect than MIR and temperature. Line 309. Given the outlined conclusions, I do not see how the results of this model enhance the ability to predict WNV incidence. Please provide more details on how this study could develop or modify currently existing thresholds for public health. Although the LightGBM method does not show improved predictive power over the linear model, its predictive power is similar and it makes those predictions using a different modeling process. This makes LightGBM suitable for inclusion in model ensembles, which benefit from inclusion of models built on different frameworks and underlying assumptions. Attachments Attachment Submitted filename: Response to Reviewers.docx https://doi.org/10.1371/journal.pone.0296283.r002
10 Dec 2023 Decision Letter - José Ramos-Castañeda, Editor Two-Step Light Gradient Boosted Model to identify human West Nile Virus infection risk factor in Chicago PONE-D-23-17817R1 Dear Dr. Ge, We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication. An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org. If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. Kind regards, José Ramos-Castañeda, M.Sc., Ph.D Academic Editor PLOS ONE Additional Editor Comments (optional): Thank you for having responded to the reviewers' comments and suggestions. One reviewer has made editorial comments on your manuscript, so the article would benefit in clarity if they were carried out. This can be done at the post-acceptance revision stage. Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation. Reviewer #1: All comments have been addressed Reviewer #2: (No Response) ******** 2. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes Reviewer #2: Yes ****** 3. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes Reviewer #2: Yes ****** 4. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes Reviewer #2: Yes ****** 5. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: Yes Reviewer #2: Yes ****** 6. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: The author justified the observations mentioned in the review of both reviewers 1 and 2. Therefore, it is suggested that we proceed to publication. Reviewer #2: Thank you for addressing my previous comments and suggestions. I believe that further work is needed to improve the clarity of the manuscript such that these findings could be meaningful for public health and vector control. In particular, the manuscript would benefit from another review by the authors with an eye towards grammar and removing redundant sentences. Also, the Conclusions are somewhat hard to follow and could be linked back to the original goals of the project for clarity for the readers. It is unclear how or why lightGBM models like yours could improve WNV prediction. I illustrate a few places with redundant sentences below: Lines 133-134. The statement “KS test is distribution-free” is redundant to “it doesn’t rely on distribution assumptions”. Lines 145-148 are restatements of Lines 140-143. Other comments: Figures 5-7. Please provide information in the legends on the significance of the green line and the faded blue lines in each partial dependence plot so the reader can understand what you are presenting here. Line 333. Misspelling of “Keyel”. Lines 342-345. This paragraph is hard to follow. How does the similar performance of the full and reduced models (first part of the paragraph) relate to variation in covariate data (second part)? ****** 7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No Reviewer #2: No ******** https://doi.org/10.1371/journal.pone.0296283.r003
Formally Accepted
28 Dec 2023 Acceptance Letter - José Ramos-Castañeda, Editor PONE-D-23-17817R1 PLOS ONE Dear Dr. Ge, I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now being handed over to our production team. At this stage, our production department will prepare your paper for publication. This includes ensuring the following: * All references, tables, and figures are properly cited * All relevant supporting information is included in the manuscript submission, * There are no issues that prevent the paper from being properly typeset If revisions are needed, the production department will contact you directly to resolve them. If no revisions are needed, you will receive an email when the publication date has been set. At this time, we do not offer pre-publication proofs to authors during production of the accepted work. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few weeks to review your paper and let you know the next and final steps. Lastly, if your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. If we can help with anything else, please email us at customercare@plos.org. Thank you for submitting your work to PLOS ONE and supporting open access. Kind regards, PLOS ONE Editorial Office Staff on behalf of Dr. José Ramos-Castañeda Academic Editor PLOS ONE https://doi.org/10.1371/journal.pone.0296283.r004

Open letter on the publication of peer review reports

PLOS recognizes the benefits of transparency in the peer review process. Therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. Reviewers remain anonymous, unless they choose to reveal their names.

We encourage other journals to join us in this initiative. We hope that our action inspires the community, including researchers, research funders, and research institutions, to recognize the benefits of published peer review reports for all parts of the research system.

Learn more at ASAPbio .