[논문]3D 프린팅 소재 화학물질의 독성 예측을 위한 Data-centric XAI 기반 분자 구조 Data Imputation과 QSAR 모델 개발

정찬혁; 김상윤; 허성구; 신민혁; 유창규

doi:10.9713/kcer.2023.61.4.523

3D 프린팅 소재 화학물질의 독성 예측을 위한 Data-centric XAI 기반 분자 구조 Data Imputation과 QSAR 모델 개발
Data-centric XAI-driven Data Imputation of Molecular Structure and QSAR Model for Toxicity Prediction of 3D Printing Chemicals 원문보기

Korean chemical engineering research = 화학공학, v.61 no.4, 2023년, pp.523 - 541

정찬혁 (경희대학교 공과대학 환경응용과학과 융합공학전공) , 김상윤 (경희대학교 공과대학 환경응용과학과 융합공학전공) , 허성구 (경희대학교 공과대학 환경응용과학과 융합공학전공) , (경희대학교 공과대학 환경응용과학과 융합공학전공) , 신민혁 (경희대학교 공과대학 환경응용과학과 융합공학전공) , 유창규 (경희대학교 공과대학 환경응용과학과 융합공학전공)

초록
AI-Helper

3D 프린터의 활용이 높아짐에 따라 발생하는 화학물질에 대한 노출 빈도가 증가하고 있다. 그러나 3D 프린팅 발생 화학물질의 독성 및 유해성에 대한 연구는 미비하며, 분자 구조 데이터의 결측치로 인해 in silico 기법을 사용한 독성예측 연구는 저조한 실정이다. 본 연구에서는 화학물질의 분자구조 정보를 나타내는 주요 분자표현자의 결측치를 보간하여 3D 프린팅의 독성 및 유해성을 예측한 Data-centric QSAR 모델을 개발하였다. 먼저 MissForest 알고리즘을 사용해 3D 프린팅으로 발생되는 유해물질의 분자표현자 결측치를 보완하였으며, 서로 다른 4가지 기계학습 모델(결정트리, 랜덤포레스트, XGBoost, SVM)을 기반으로 Data-centric QSAR 모델을 개발하여 생물 농축 계수(Log BCF)와 옥탄올-공기분배계수(Log Koa), 분배계수(Log P)를 예측하였다. 또한, 설명 가능한 인공지능(XAI) 방법론 중 TreeSHAP (SHapley Additive exPlanations) 기법을 활용하여 Data-centric QSAR 모델의 신뢰성을 입증하였다. MissForest 알고리즘 기반 결측지 보간 기법은, 기존 분자구조 데이터에 비하여 약 2.5배 많은 분자구조 데이터를 확보할 수 있었다. 이를 바탕으로 개발된 Data-centric QSAR 모델의 성능은 Log BCF, Log Koa와 Log P를 각각 73%, 76%, 92% 의 예측 성능으로 예측할 수 있었다. 마지막으로 Tree-SHAP 분석결과 개발된 Data-centric QSAR 모델은 각 독성치와 물리적으로 상관성이 높은 분자표현자를 통하여 선택함을 설명할 수 있었고 독성 정보에 대한 높은 예측 성능을 확보할 수 있었다. 본 연구에서 개발한 방법론은 다른 프린팅 소재나 화학공정, 그리고 반도체/디스플레이 공정에서 발생 가능한 오염물질의 독성 및 인체 위해성 평가에 활용될 수 있을 것으로 사료된다.

Abstract ▼ AI-Helper

As accessibility to 3D printers increases, there is a growing frequency of exposure to chemicals associated with 3D printing. However, research on the toxicity and harmfulness of chemicals generated by 3D printing is insufficient, and the performance of toxicity prediction using in silico techniques is limited due to missing molecular structure data. In this study, quantitative structure-activity relationship (QSAR) model based on data-centric AI approach was developed to predict the toxicity of new 3D printing materials by imputing missing values in molecular descriptors. First, MissForest algorithm was utilized to impute missing values in molecular descriptors of hazardous 3D printing materials. Then, based on four different machine learning models (decision tree, random forest, XGBoost, SVM), a machine learning (ML)-based QSAR model was developed to predict the bioconcentration factor (Log BCF), octanol-air partition coefficient (Log Koa), and partition coefficient (Log P). Furthermore, the reliability of the data-centric QSAR model was validated through the Tree-SHAP (SHapley Additive exPlanations) method, which is one of explainable artificial intelligence (XAI) techniques. The proposed imputation method based on the MissForest enlarged approximately 2.5 times more molecular structure data compared to the existing data. Based on the imputed dataset of molecular descriptor, the developed data-centric QSAR model achieved approximately 73%, 76% and 92% of prediction performance for Log BCF, Log Koa, and Log P, respectively. Lastly, Tree-SHAP analysis demonstrated that the data-centric-based QSAR model achieved high prediction performance for toxicity information by identifying key molecular descriptors highly correlated with toxicity indices. Therefore, the proposed QSAR model based on the data-centric XAI approach can be extended to predict the toxicity of potential pollutants in emerging printing chemicals, chemical process, semiconductor or display process.

주제어

표/그림 (19)

그림 Fig. 1. Hazardous pollutants emitted from 3D printing processes according to filaments including Acrylonitrile butadiene styrene (ABS), Polylactic acid (PLA), High impact polystyrene (HIPs), Nylon, Polycarbonate (PC) and Polypropylene (PP).
표 Table 1. Summary of previous studies related to 3D printing, QSAR, and missing data imputation
표 Table 2. Monomer structure of 3D printing filaments and hazardous pollutants emitted from 3D printing operation
그림 Fig. 2. Research framework of molecular descriptors imputation and data-centric XAI-driven QSAR model for toxicity prediction of 3D printing chemicals.
표 Table 3. The molecular descriptors in each group calculated by DRAGON 6 software [15]
표 Table 4. Toxicity indices and the number of 3D printing chemicals and percentage of those missing molecular data
그림 Fig. 3. Balance of actual and missing molecular descriptors in missing-data-included groups.
그림 Fig. 4. MissForest algorithm workflow for the missing imputation of molecular descriptors.
그림 Fig. 5. Overall workflow of ML-QSAR model development: (a) Structure of machine learning models, (b) Model training procedure, and (c) Hyperparameter tuning using grid search.
그림 Fig. 6. Training result of MissForest algorithm for missing molecular descriptor data imputation: (a) Training dataset (b) Validation dataset, and (c) Test dataset.
그림 Fig. 7. Distribution of (a) original and (b) imputed molecular descriptor data, and (c) imputation results for molecular descriptors in 2D matrix-based descriptor group by MissForest algorithm.
그림 Fig. 8. Comparison of the number of molecular descriptors between imputed and non-imputed molecular descriptors dataset.
표 Table 5. Hyperparameters tuning results of the data-centric QSAR model by GridSearch
표 Table 6. Comparison of prediction performances in data-centric and model-centric QSAR model
그림 Fig. 9. R2 score comparisons of the data-centric and model-centric QSAR models across toxicity indices.
그림 Fig. 10. Prediction results using ML-QASR models based on XGBoost and SVM for (a) Log BCF, (b) log KOA, and (c) Log P.
그림 Fig. 11. Comparison of local explanation diagram of SHAP values for data-centric and model-centric QSAR models of (a) Log BCF (b) Log Koa, and (c) Log P.
그림 Fig. 12. Waterfall SHAP plots for (a) high Log Koa, (b) low Log Koa, (c) low Log P, and (d) high Log P 3D printing chemicals.
표 Table 7. Summary of the data-centric and model-centric QSAR model for toxicity prediction of 3D printing chemicals

참고문헌 (39)

Kwon, K. M., Kim, H. G. and Moon, S. H., "The 4th Basic Plan？for Material and Parts Development," Ministry of Trade, Industry and Energy 16-28(2016).
Jang, Y. J. and Jeong, E. M.,, "Global Trends in the 4th Industrial Revolution and Strategies for the Response of Korean？Industries," Korea Institute for Industrial Economics & Trade？22-24(2017).
Park, S. H., "A Study on R&D Policy through 3D Printing？Industry Trend Analysis," Science and Technology Policy 24(3),？93-104(2014).
An, K. C., "Trends and Implications of 3D Printing Industry in？the 4th Industrial Revolution," Institute for Information & Communications Technology Promotion 5-8(2018).
Zhou, Y., Kong, X., Chen, A. and Cao, S., "Investigation of Ultrafine Particle Emissions of Desktop 3D Printers in the Clean？Room," Procedia Eng., 121, 506-512(2015).
HUBS, Additive Manufacturing Trend Report 2021 (2021).
Stabile, L., Scungio, M., Buonanno, G., Arpino, F. and Ficco,？G., "Airborne Particle Emission of a Commercial 3D Printer: the？Effect of Filament Material and Printing Temperature," Indoor？Air, 27(2), 398-408(2017).

상세보기
Kim, Y. N., Yoon, C. S., Ham, S. H., Park, J. H., Kim, S. H.,？Kwon, O. H. and Tsai, P. J., "Emissions of Nanoparticles and？Gaseous Material from 3D Printer Operation," Environ Sci Technol., 49(20), 12044-12053(2015).

상세보기
Azimi, P., Zhao, D., Pouzet, C., Crain, N. E. and Stephens B.,？"Emissions of Ultrafine Particles and Volatile Organic Compounds from Commercially Available Desktop Three-Dimensional？Printers with Multiple Filaments," Environ Sci Technol., 50(3),？1260-1268(2016).

상세보기
Steinle, P., "Characterization of Emissions from a Desktop 3D？Printer and Indoor Air Measurements in Office Settings," J.？Occup. Environ. Hyg., 13(2), 121-132(2016).

상세보기
Kim G. H., Lyu K. G., Kim Y. J. and Kim H. C., "A Survey on？Quantitative Structure-Activity Relationship(QSAR) Models,"？2008 Proceedings of the Korean Information Science Society？Conference, July, Pyungchang 35(1), 43-44(2008).
Ding Y. L., Lyu Y. C. and Leong M. K., "In Silico Prediction of？the Mutagenicity of Nitroaromatic Compounds Using a Novel？Two-QSAR Approach," Toxicology in Vitro 40(1), 102-114(2017).
Kobayashi, Y. and Yoshida, K., "Development of QSAR Models for Prediction of Fish Bioconcentration Factors Using Physicochemical Properties and Molecular Descriptors with Machine？Learning Algorithms," Ecol Inform 63(1), 2-9(2021).
Pandit, S., Singh, P., Sinha, M. and Parthasarathi, R., "Integrated？QSAR and Adverse Outcome Pathway Analysis of Chemicals？Released on 3D Printing Using Acrylonitrile Butadiene Styrene," Chem Res Toxicol 34(2), 355-364(2021).

상세보기
Kim D. W., Lee S. C., Kim M. J., Lee E. J. and Yoo C. K.,？"Development of QSAR Model Based on the Key Molecular？Descriptors Selection and Computational Toxicology for Prediction of Toxicity of PCBs," Korean Chemical Engineering Research？54(5), 621-629(2016).
To, K. T., Fry, R. C. and Reif, D. M., "Characterizing the Effects？of Missing Data and Evaluating Imputation Methods for Chemical？Prioritization Applications Using ToxPi," BioData Min 11(1),？(2018).

상세보기
Lee J. G., Shin G. J., Park C. Y. and Hwang U. J., "Robust, fair？and scalable data-driven continuous learning," Communications？of the Korean Institute of Information Scientists and Engineers？40(11), 53-58(2022).
Yang, F., Du, J., Lang, J., Lu, W., Liu, L., Jin, C. and Kang, Q.,？"Missing Value Estimation Methods Research for Arrhythmia？Classification Using the Modified Kernel Difference-Weighted？KNN Algorithms," Biomed Res Int 2020(1), 1-9(2020).
Luo, Y., "Evaluating the State of the Art in Missing Data Imputation for Clinical Data," Brief Bioinform 23(1), 1-9(2022).
Carli, M., Ward, M. H., Metayer, C. and Wheeler, D. C., "Imputation of Below Detection Limit Missing Data in Chemical Mixture Analysis with Bayesian Group Index Regression," Int. J.？Environ Res. Public Health 19(3), 2-14(2022).
Jeong, J. S., Garcia-Reyero, N., Burgoon, L., Perkins, E., Park,？T. H., Kim, C. H., Roh, J. Y. and Choi, J. H., "Development of？Adverse Outcome Pathway for PPARγ Antagonism Leading to？Pulmonary Fibrosis and Chemical Selection for Its Validation:？ToxCast Database and a Deep Learning Artificial Neural Network Model-Based Approach," Chem Res Toxicol 32(6), 1212-1222(2019).
Tiganis, B. E., Burn, L. S., Davis, P. and Hill, A. J., "Thermal？Degradation of Acrylonitrile-butadiene-styrene (ABS) Blends,"？Polym Degrad Stab 76(1), 425-434(2002).
Rutkowski, J. V. and Levin, B. C., "Acrylonitrile-Butadiene-Styrene Copolymers (ABS): Pyrolysis and Combustion Products？and their Toxicity-A Review of the Literature," Fire Mater 10(1),？93-105(1986).
Wojtyla, S., Klama, P. and Baran, T., "Is 3D Printing Safe? Analysis of the Thermal Treatment of Thermoplastics: ABS, PLA,？PET, and Nylon," J. Occup. Environ. Hyg., 14(6), 80-85(2017).
Davis, A. Y., Zhang, Q., Wong, J. P. S., Weber, R. J. and Black,？M. S., "Characterization of Volatile Organic Compound Emissions from Consumer Level Material Extrusion 3D Printers,"？Build Environ 160, 106209(2019).

상세보기
Pandit, S., Singh, P., Sinha, M. and Parthasarathi, R., "Integrated？QSAR and Adverse Outcome Pathway Analysis of Chemicals？Released on 3D Printing Using Acrylonitrile Butadiene Styrene," Chem Res Toxicol 34(2), 355-364(2021).

상세보기
Park, J. H., Jeon, H. J., Oh, Y. S., Park, K. H. and Yoon, C. S.,？"Understanding Three-dimensional Printing Technology, Evaluation, and Control of Hazardous Exposure Agents," Journal of？Korean Society of Occupational and Environmental Hygiene？28(3), 241-256(2018).
Kim, S. H., Chung, E. K., Kim, S. D. and Kwon, J. W., "Assessment of Emitted Volatile Organic Compounds, Metals and Characteristic of Particle in Commercial 3D Printing Service？Workplace," Original Article Journal of Korean Society of Occupational and Environmental Hygiene 30(2), 153-162(2020).
Kim, S. H. and Chung, E. K., "A Study on the Types of Materials？and Hazardous Substances Used in 3D Printers," Korea Occupational Safety and Health Agency 10-13(2019).
Hong, M. K., Jo, J. H., Choi, B. K. and Kim, K. W., A Study on？the Application of OECD Toolbox in Chemical Information (2018).
Mauri, A., Srl, A., Consonni, V., Pavan, M. and Todeschini, R.,？DRAGON SOFTWARE: AN EASY APPROACH TO MOLECULAR DESCRIPTOR CALCULATIONS (n.d.).
Rahimi, R., Keshavarz, M. H., and Akbarzadeh, A. R., "Prediction of the Density of Energetic Materials on the Basis of their？Molecular Structures," Central European Journal of Energetic？Materials 13(1), 73-101(2016).
Consonni, V., Todeschini, R. and Pavan, M., "Structure/response？Correlations and Similarity/diversity Analysis by GETAWAY Descriptors. 1. Theory of the Novel 3D Molecular Descriptors," J. Chem.？Inf. Comput. Sci., 42(3), 682-692(2002).

상세보기
Devinyak, O., Havrylyuk, D. and Lesyk, R., "3D-MoRSE Descriptors Explained," J. Mol. Graph. Model., 54, 194-203(2014).

상세보기
Stekhoven, D. J. and Buhlmann, P., "Missforest-Non-parametric？Missing Value Imputation for Mixed-type Data," Bioinformatics？28(1), 112-118(2012).

상세보기
Marinov, D. and Karapetyan, D., "Hyperparameter Optimisation？with Early Termination of Poor Performers," 2019 11th Computer？Science and Electronic Engineering (CEEC), Colchester, UK？160-163(2019).
Choi, G. C., Kim, W. J. and Koo, J. M., "Investigating the Performance of Machine Learning Methods in Predicting Functional Properties of the Hydrogenase Variants," Biotechnology？and Bioprocess Engineering 28(1), 143-151(2023).

상세보기
Moon, J. H., Park, S. W., Rho, S. M. and Hwang, E. J., "Interpretable Short-Term Electrical Load Forecasting Scheme Using？Cubist," Comput Intell Neurosci 2022(1), 2-19(2022).
Lundberg, S. M., Allen, P. G. and Lee, S. I., "A Unified Approach？to Interpreting Model Predictions," 31st Conference on Neural？Information Processing Systems (NIPS 2017), December, California 1-10(2017).

저자의 다른 논문 :

표제어: PCR

동의어: Packet Collision Rate

용어 설명 출처 목록 (6)

용어 설명: PCR은 세균 특이성이 있는 primer를 이용하여 적은 수의 세균이 있을지라도 쉽게 검출할 수 있는 유용한 방법이며, 이를 이용하여 구강 내 치면세균막이나 타액에서 직접 세균을 검출할 수 있게 되었다[8].

내보내기 구분	파일저장 인쇄 메일전송
구성항목	기본정보 상세정보 관리번호, 논문명, 저널/프로시딩명, 저자 , 발행년, 권, 호, 시작페이지, 끝페이지, 발행기관 관리번호, 논문명, 대등논문명, 저자 , 저널/프로시딩명, 발행기관, 발행년, 발행언어, 권, 호, 시작페이지, 끝페이지, ISBN, ISSN, 주제분야, 키워드, 초록(한글), 초록(영문), 저자(소속기관)
저장형식	Text(ASCII format) Excel format RefWorks Direct Export RIS format (for Reference Manager, ProCite, EndNote), Scholar's Aids, Mendeley
메일정보	받는사람 (필수) @ 보내는사람 (선택) @ 제목 내용 KISTI 검색결과 이메일 서비스
안내	총 건의 자료가 검색되었습니다. 다운받으실 자료의 인덱스를 입력하세요. (1-10,000) 검색결과의 순서대로 최대 10,000건 까지 다운로드가 가능합니다. 데이타가 많을 경우 속도가 느려질 수 있습니다.(최대 2~3분 소요) 다운로드 파일은 UTF-8 형태로 저장됩니다. 파일의 내용이 제대로 보이지 않을실 때는 웹브라우저 상단의 보기 -> 인코딩 -> 자동선택 여부를 확인하십시오. ~ Text(ASCII format) Excel format

연합인증