Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice letter Brief Bioinform . 2026 Apr 8;27(2):bbag161. doi: 10.1093/bib/bbag161 Search in PMC Search in PubMed View in NLM Catalog Add to search Response to “Addressing biases and limitations in feature attribution for circRNA modification profiling” Jiayi Li Jiayi Li 1 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 2 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 3 Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool L69 7ZB, United Kingdom Find articles by Jiayi Li 1, 2, 3 , Shenglun Chen Shenglun Chen 4 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 5 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 6 Institute of Infection, Veterinary and Ecological Sciences, University of Liverpool, Leahurst Campus, Neston, Brownlow Hill, Wirral CH64 7TE, United Kingdom Find articles by Shenglun Chen 4, 5, 6 , Zhixing Wu Zhixing Wu 7 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 8 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 9 The School of AI and Advanced Computing, XJTLU Entrepreneur College (Taicang), Xi'an Jiaotong-Liverpool University, 5th Floor, Building D, No. 111 Road, Taicang Avenue, Taicang, Suzhou, Jiangsu 215400, China 10 Department of Computer Science, University of Liverpool, Ashton Building, Liverpool, Merseyside L69 3DR, United Kingdom Find articles by Zhixing Wu 7, 8, 9, 10 , Haozhe Wang Haozhe Wang 11 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 12 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 13 Department of Computer Science, University of Liverpool, Ashton Building, Liverpool, Merseyside L69 3DR, United Kingdom Find articles by Haozhe Wang 11, 12, 13 , Rong Xia Rong Xia 14 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 15 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 16 The School of AI and Advanced Computing, XJTLU Entrepreneur College (Taicang), Xi'an Jiaotong-Liverpool University, 5th Floor, Building D, No. 111 Road, Taicang Avenue, Taicang, Suzhou, Jiangsu 215400, China 17 Department of Computer Science, University of Liverpool, Ashton Building, Liverpool, Merseyside L69 3DR, United Kingdom Find articles by Rong Xia 14, 15, 16, 17 , Jia Meng Jia Meng 18 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 19 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 20 Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool L69 7ZB, United Kingdom Find articles by Jia Meng 18, 19, 20 , Yuxin Zhang Yuxin Zhang 21 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 22 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 23 Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool L69 7ZB, United Kingdom 24 State Key Laboratory of Epigenetic Regulation and Intervention, Institute of Biophysics, Chinese Academy of Sciences, No. 15 Datun Road, Chaoyang District, Beijing 100101, China Find articles by Yuxin Zhang 21, 22, 23, 24, ✉ Author information Article notes Copyright and License information 1 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 2 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 3 Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool L69 7ZB, United Kingdom 4 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 5 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 6 Institute of Infection, Veterinary and Ecological Sciences, University of Liverpool, Leahurst Campus, Neston, Brownlow Hill, Wirral CH64 7TE, United Kingdom 7 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 8 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 9 The School of AI and Advanced Computing, XJTLU Entrepreneur College (Taicang), Xi'an Jiaotong-Liverpool University, 5th Floor, Building D, No. 111 Road, Taicang Avenue, Taicang, Suzhou, Jiangsu 215400, China 10 Department of Computer Science, University of Liverpool, Ashton Building, Liverpool, Merseyside L69 3DR, United Kingdom 11 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 12 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 13 Department of Computer Science, University of Liverpool, Ashton Building, Liverpool, Merseyside L69 3DR, United Kingdom 14 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 15 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 16 The School of AI and Advanced Computing, XJTLU Entrepreneur College (Taicang), Xi'an Jiaotong-Liverpool University, 5th Floor, Building D, No. 111 Road, Taicang Avenue, Taicang, Suzhou, Jiangsu 215400, China 17 Department of Computer Science, University of Liverpool, Ashton Building, Liverpool, Merseyside L69 3DR, United Kingdom 18 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 19 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 20 Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool L69 7ZB, United Kingdom 21 Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 22 Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China 23 Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool L69 7ZB, United Kingdom 24 State Key Laboratory of Epigenetic Regulation and Intervention, Institute of Biophysics, Chinese Academy of Sciences, No. 15 Datun Road, Chaoyang District, Beijing 100101, China ✉ Corresponding author. Department of Biosciences and Bioinformatics, School of Science, Xi’an Jiaotong–Liverpool University, Suzhou, Jiangsu 215123, China. E-mail: [email protected] . Received 2026 Mar 8; Accepted 2026 Mar 17; Collection date 2026 Mar. © The Author(s) 2026. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( https://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. PMC Copyright notice PMCID: PMC13069882 PMID: 41955025 Abstract This response addresses the comments raised by Souichi Oka and colleagues in their Letter to the Editor titled “Addressing biases and limitations in feature attribution for circRNA modification profiling.” We clarify that two independent XGBoost models were used for distinct purposes in our analysis: one for predicting RNA modification events from nanopore-derived signal features and another for feature attribution using genome-derived sequence features extracted through the m6AlogisticModel framework. We further note that Shapley Additive Explanations (SHAP) was employed as an exploratory interpretability tool rather than as definitive evidence of causal biological mechanisms. We appreciate the constructive methodological suggestions provided and acknowledge that integrating complementary analytical strategies may further enhance the robustness of computational studies of circRNA modifications. Keywords: circRNA, feature attribution, model interpretability Dear Editor, We thank Souichi Oka and colleagues for their thoughtful comments on our article, CircRM: profiling circular RNA modifications from nanopore direct RNA sequencing. Their discussion raises several important issues concerning the interpretation of feature importance and the methodological limitations of machine-learning–based analyses in epitranscriptomic studies. We appreciate the opportunity to clarify several aspects of our study. First, we would like to clarify the modeling framework used in our analysis. In our study, two distinct XGBoost models were constructed for different analytical purposes. The first model, which incorporated nanopore signal-derived features such as dwell time and other signal statistics, was developed solely for the prediction of RNA modification events. In contrast, the feature attribution analysis reported in our study was performed using a second, independently trained model. For this analysis, genome-derived sequence features were extracted using the feature representation strategy implemented in m6AlogisticModel [ 1 ], after which a separate XGBoost model was trained based on these features. Importantly, this second model is independent of the prediction model based on nanopore signal features. Therefore, the SHAP-based interpretation presented in our study reflects the behavior of the sequence-feature model rather than the signal-based prediction model. Second, the use of SHAP for model interpretation represents a widely adopted approach in machine learning and bioinformatics. SHAP provides a theoretically grounded framework for estimating feature contributions based on cooperative game theory and has been widely used to interpret complex predictive models across diverse biological applications [ 2 , 3 ]. While we acknowledge that no attribution method is without limitations, SHAP remains a commonly used tool for examining how predictive features contribute to model outputs. In our study, SHAP analysis was used primarily as an exploratory approach to identify features that may be associated with the model’s predictions, rather than as definitive evidence of causal biological mechanisms. Third, we appreciate the authors’ constructive suggestions regarding potential methodological improvements. In particular, approaches such as highly variable feature selection, feature agglomeration, and complementary non-parametric statistical analyses may provide additional perspectives for evaluating feature relevance in complex biological datasets. We agree that incorporating multiple analytical perspectives can strengthen the robustness of computational inference. These suggestions are valuable and will be considered in future work aimed at further improving the analytical framework for circRNA modification profiling. In summary, we thank the authors for their constructive comments and for highlighting important considerations regarding model interpretability. We hope that the clarifications provided here help to better explain the analytical design of our study. Continued methodological discussion will undoubtedly contribute to improving computational approaches for studying RNA modifications. Key Points CircRM employs two independent XGBoost models for prediction and interpretability analyses. SHAP was applied as an exploratory interpretability tool to examine feature contributions to model predictions rather than to establish causal biological mechanisms. The methodological suggestions proposed in the Letter are appreciated and may help further improve the robustness of computational analyses for circRNA modification profiling. Contributor Information Jiayi Li, Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool L69 7ZB, United Kingdom. Shenglun Chen, Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Institute of Infection, Veterinary and Ecological Sciences, University of Liverpool, Leahurst Campus, Neston, Brownlow Hill, Wirral CH64 7TE, United Kingdom. Zhixing Wu, Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; The School of AI and Advanced Computing, XJTLU Entrepreneur College (Taicang), Xi'an Jiaotong-Liverpool University, 5th Floor, Building D, No. 111 Road, Taicang Avenue, Taicang, Suzhou, Jiangsu 215400, China; Department of Computer Science, University of Liverpool, Ashton Building, Liverpool, Merseyside L69 3DR, United Kingdom. Haozhe Wang, Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Department of Computer Science, University of Liverpool, Ashton Building, Liverpool, Merseyside L69 3DR, United Kingdom. Rong Xia, Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; The School of AI and Advanced Computing, XJTLU Entrepreneur College (Taicang), Xi'an Jiaotong-Liverpool University, 5th Floor, Building D, No. 111 Road, Taicang Avenue, Taicang, Suzhou, Jiangsu 215400, China; Department of Computer Science, University of Liverpool, Ashton Building, Liverpool, Merseyside L69 3DR, United Kingdom. Jia Meng, Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool L69 7ZB, United Kingdom. Yuxin Zhang, Department of Biosciences and Bioinformatics, School of Science, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Center for Intelligent RNA Therapeutics, Suzhou Key Laboratory of Cancer Biology and Chronic Disease, Xi'an Jiaotong-Liverpool University, 111 Ren'ai Road, Suzhou Industrial Park, Suzhou, Jiangsu 215123, China; Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool L69 7ZB, United Kingdom; State Key Laboratory of Epigenetic Regulation and Intervention, Institute of Biophysics, Chinese Academy of Sciences, No. 15 Datun Road, Chaoyang District, Beijing 100101, China. Conflict of interest None declared. Funding None declared. Data availability There are no new data associated with this article. References 1. CompGenomics-xjtlu . ZW-xjtlu/m6ALogisticModel, https://github.com/ZW-xjtlu/m6ALogisticModel . 2025. 2. Tu G, Wang X, Xia R et al. m6A-TCPred: a web server to predict tissue-conserved human m6A sites using machine learning approach. BMC Bioinformatics 2024;25:127. 10.1186/s12859-024-05738-1 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Xia R, Yin X, Huang J et al. Interpretable deep cross networks unveiled common signatures of dysregulated epitranscriptomes across 12 cancer types. Mol Ther Nucleic Acids 2024;35:102376. 10.1016/j.omtn.2024.102376 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Data Availability Statement There are no new data associated with this article. Articles from Briefings in Bioinformatics are provided here courtesy of Oxford University Press ACTIONS View on publisher site PDF (259.8 KB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top