Dat Duong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Dat Duong has a total of 64 co-authors. Here are the top co-authors by number of shared publications:
- Benjamin D. Solomon
Medical Genetics Branch, National Human Genome Research Institute, Bethesda, Maryland, United States of America
5 shared publications - Rebekah L. Waikel
Medical Genetics Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, Maryland, USA
4 shared publications Eleazar Eskin
University of California Los Angeles
3 shared publications- Eleazar Eskin
UCLA
3 shared publications - Jingyi Jessica Li
Biostatistics Program, Public Health Sciences Division, Fred Hutchinson Cancer Center, Seattle, USA
3 shared publications Lisa Gai
3 shared publications
Latest Publications
- 23 September 2026
- Published by Wiley in Journal of Genetic Counseling
- Vol. 35 (5) , e70296
Artificial Intelligence‐Based Chatbots in Genetic Counseling Practice: Current Uptake, Utilization, and Perspectives
AI‐driven chatbots have been utilized in healthcare to automate administrative tasks, improve patient education, and expand access to medical information; however, their role in genetic counseling remains underexplored. To investigate the adoption, perceptions, and potential utility of AI‐based chatbots in genetic counseling practice, 217 genetic counselors and genetic counseling students from across North America were surveyed regarding chatbot usage, confidence in their application, and perceived benefits and limitations. While most participants (166/217; 76.5%) reported using general AI chatbots outside of clinical settings, far fewer (18/204; 8.8%) reported using or recommending clinical genetics chatbots in clinical practice. For those that used clinical genetics chatbots, the primary purpose was for communication with at‐risk family members (11/18; 61.1%) and patient education (10/18; 55.6%). Confidence in chatbot technology varied, with highest confidence in gathering family history information (81/199; 40.7%) and lowest confidence in their ability to disclose variants of uncertain significance or positive genetic testing results (5/199; 2.5%). The greatest perceived benefits included reducing repetitive tasks (165/195, 84.6%) and allowing for time for other tasks (141/195; 72.3%), while major concerns revolved around patient comprehension (167/195; 85.6%) and having accurate, up‐to‐date information (145/195; 74.4%). Despite some concern about AI replacing human counselors, most participants reported they felt there was potential for chatbots to enhance workflow efficiency (128/195; 65.6%) if properly integrated and regulated. Limited AI training was identified as a barrier to adoption (16/195; 8.2% received training), highlighting a need for structured education on AI applications in genetic counseling. These findings suggest that AI chatbots hold promise as supplementary tools, but significant challenges must be addressed before widespread implementation in genetic counseling practice.
- 23 September 2026
- Published by Wiley in Journal of Genetic Counseling
- Vol. 35 (5) , e70296
Artificial Intelligence‐Based Chatbots in Genetic Counseling Practice: Current Uptake, Utilization, and Perspectives
AI‐driven chatbots have been utilized in healthcare to automate administrative tasks, improve patient education, and expand access to medical information; however, their role in genetic counseling remains underexplored. To investigate the adoption, perceptions, and potential utility of AI‐based chatbots in genetic counseling practice, 217 genetic counselors and genetic counseling students from across North America were surveyed regarding chatbot usage, confidence in their application, and perceived benefits and limitations. While most participants (166/217; 76.5%) reported using general AI chatbots outside of clinical settings, far fewer (18/204; 8.8%) reported using or recommending clinical genetics chatbots in clinical practice. For those that used clinical genetics chatbots, the primary purpose was for communication with at‐risk family members (11/18; 61.1%) and patient education (10/18; 55.6%). Confidence in chatbot technology varied, with highest confidence in gathering family history information (81/199; 40.7%) and lowest confidence in their ability to disclose variants of uncertain significance or positive genetic testing results (5/199; 2.5%). The greatest perceived benefits included reducing repetitive tasks (165/195, 84.6%) and allowing for time for other tasks (141/195; 72.3%), while major concerns revolved around patient comprehension (167/195; 85.6%) and having accurate, up‐to‐date information (145/195; 74.4%). Despite some concern about AI replacing human counselors, most participants reported they felt there was potential for chatbots to enhance workflow efficiency (128/195; 65.6%) if properly integrated and regulated. Limited AI training was identified as a barrier to adoption (16/195; 8.2% received training), highlighting a need for structured education on AI applications in genetic counseling. These findings suggest that AI chatbots hold promise as supplementary tools, but significant challenges must be addressed before widespread implementation in genetic counseling practice.
- 9 July 2026
- Published by Elsevier in Genetics in Medicine
- Vol. 28 (9) , 102645
Artificial intelligence in clinical genetics: current practice and attitudes among the clinical genetics workforce
- 9 July 2026
- Published by Elsevier in Genetics in Medicine
- Vol. 28 (9) , 102645
Artificial intelligence in clinical genetics: current practice and attitudes among the clinical genetics workforce
- Open Access
Artificial intelligence in clinical genetics: current practice and attitudes among the clinical genetics workforce
Purpose: Artificial intelligence (AI) applications for clinical genetics hold the potential to improve patient care through supporting diagnostics and management as well as automating administrative tasks, thus enhancing and potentially enabling clinician/patient interactions. While the introduction of AI into clinical genetics is increasing, there remain unclear questions about risks and benefits, and the readiness of the workforce. Methods: To assess the current clinical genetics workforce’s use, knowledge, and attitudes toward available medical AI applications, we conducted a survey involving 215 US-based genetics clinicians and trainees. Results: Over half (51.2%) of participants report little to no knowledge of AI in clinical genetics and 64.3% reported no formal training in AI applications. Formal training directly correlated with self-reported knowledge of AI in clinical genetics, with 69.3% of respondents with formal training reporting intermediate to extensive knowledge of AI vs. 37.5% without formal training. Most participants reported that they lacked sufficient knowledge of clinical AI (83.4%) and agreed that there should be more education in this area (97.6%) and would take a course if offered (89.3%). The majority (51.6%) of clinician participants said they never used AI applications in the clinic. However, after a tutorial describing clinical AI applications, 75.8% reported some use of AI applications in the clinic. When asked specifically about clinical AI application usage, the majority of clinician participants used facial diagnostic applications (54.9%) and AI-generated genomic testing results (62.1%), whereas other applications such as chatbots, large language models (LLMs), pedigree or medical summary generators, and risk assessment were only used by a fraction of the clinicians, ranging from 11.1 to 12.5%. Nearly all participants (94.6%) reported clinical genetics professionals as being overburdened. Conclusion: Further clinician education is both desired and needed to optimally utilize clinical AI applications with the potential to enhance patient care and alleviate the current strain on genetics clinics.
Open AccessArtificial intelligence in clinical genetics: current practice and attitudes among the clinical genetics workforce
Purpose: Artificial intelligence (AI) applications for clinical genetics hold the potential to improve patient care through supporting diagnostics and management as well as automating administrative tasks, thus enhancing and potentially enabling clinician/patient interactions. While the introduction of AI into clinical genetics is increasing, there remain unclear questions about risks and benefits, and the readiness of the workforce. Methods: To assess the current clinical genetics workforce’s use, knowledge, and attitudes toward available medical AI applications, we conducted a survey involving 215 US-based genetics clinicians and trainees. Results: Over half (51.2%) of participants report little to no knowledge of AI in clinical genetics and 64.3% reported no formal training in AI applications. Formal training directly correlated with self-reported knowledge of AI in clinical genetics, with 69.3% of respondents with formal training reporting intermediate to extensive knowledge of AI vs. 37.5% without formal training. Most participants reported that they lacked sufficient knowledge of clinical AI (83.4%) and agreed that there should be more education in this area (97.6%) and would take a course if offered (89.3%). The majority (51.6%) of clinician participants said they never used AI applications in the clinic. However, after a tutorial describing clinical AI applications, 75.8% reported some use of AI applications in the clinic. When asked specifically about clinical AI application usage, the majority of clinician participants used facial diagnostic applications (54.9%) and AI-generated genomic testing results (62.1%), whereas other applications such as chatbots, large language models (LLMs), pedigree or medical summary generators, and risk assessment were only used by a fraction of the clinicians, ranging from 11.1 to 12.5%. Nearly all participants (94.6%) reported clinical genetics professionals as being overburdened. Conclusion: Further clinician education is both desired and needed to optimally utilize clinical AI applications with the potential to enhance patient care and alleviate the current strain on genetics clinics.
- Open Access
- 27 February 2024
- Published by Public Library of Science (PLoS) in PLOS Genetics
- Vol. 20 (2) , e1011168
Comparison of clinical geneticist and computer visual attention in assessing genetic conditions
Artificial intelligence (AI) for facial diagnostics is increasingly used in the genetics clinic to evaluate patients with potential genetic conditions. Current approaches focus on one type of AI called Deep Learning (DL). While DL- based facial diagnostic platforms have a high accuracy rate for many conditions, less is understood about how this technology assesses and classifies (categorizes) images, and how this compares to humans. To compare human and computer attention, we performed eye-tracking analyses of geneticist clinicians (n = 22) and non-clinicians (n = 22) who viewed images of people with 10 different genetic conditions, as well as images of unaffected individuals. We calculated the Intersection-over-Union (IoU) and Kullback–Leibler divergence (KL) to compare the visual attentions of the two participant groups, and then the clinician group against the saliency maps of our deep learning classifier. We found that human visual attention differs greatly from DL model’s saliency results. Averaging over all the test images, IoU and KL metric for the successful (accurate) clinician visual attentions versus the saliency maps were 0.15 and 11.15, respectively. Individuals also tend to have a specific pattern of image inspection, and clinicians demonstrate different visual attention patterns than non-clinicians (IoU and KL of clinicians versus non-clinicians were 0.47 and 2.73, respectively). This study shows that humans (at different levels of expertise) and a computer vision model examine images differently. Understanding these differences can improve the design and use of AI tools, and lead to more meaningful interactions between clinicians and AI technologies.
- 27 February 2024
- Published by Public Library of Science (PLoS) in PLOS Genetics
- Vol. 20 (2) , e1011168
Open AccessComparison of clinical geneticist and computer visual attention in assessing genetic conditions
Artificial intelligence (AI) for facial diagnostics is increasingly used in the genetics clinic to evaluate patients with potential genetic conditions. Current approaches focus on one type of AI called Deep Learning (DL). While DL- based facial diagnostic platforms have a high accuracy rate for many conditions, less is understood about how this technology assesses and classifies (categorizes) images, and how this compares to humans. To compare human and computer attention, we performed eye-tracking analyses of geneticist clinicians (n = 22) and non-clinicians (n = 22) who viewed images of people with 10 different genetic conditions, as well as images of unaffected individuals. We calculated the Intersection-over-Union (IoU) and Kullback–Leibler divergence (KL) to compare the visual attentions of the two participant groups, and then the clinician group against the saliency maps of our deep learning classifier. We found that human visual attention differs greatly from DL model’s saliency results. Averaging over all the test images, IoU and KL metric for the successful (accurate) clinician visual attentions versus the saliency maps were 0.15 and 11.15, respectively. Individuals also tend to have a specific pattern of image inspection, and clinicians demonstrate different visual attention patterns than non-clinicians (IoU and KL of clinicians versus non-clinicians were 0.47 and 2.73, respectively). This study shows that humans (at different levels of expertise) and a computer vision model examine images differently. Understanding these differences can improve the design and use of AI tools, and lead to more meaningful interactions between clinicians and AI technologies.
- 15 September 2023
- Published by Elsevier
- p. 329-339
Artificial intelligence in medical genetics
- Open Access
- 2 February 2020
- Published by Cold Spring Harbor Laboratory in bioRxiv
Annotating Gene Ontology terms for protein sequences with the Transformer model
Predicting functions for novel amino acid sequences is a long-standing research problem. The Uniprot database which contains protein sequences annotated with Gene Ontology (GO) terms, is one commonly used training dataset for this problem. Predicting protein functions can then be viewed as a multi-label classification problem where the input is an amino acid sequence and the output is a set of GO terms. Recently, deep convolutional neural network (CNN) models have been introduced to annotate GO terms for protein sequences. However, the CNN architecture can only model close-range interactions between amino acids in a sequence. In this paper, first, we build a novel GO annotation model based on the Transformer neural network. Unlike the CNN architecture, the Transformer models all pairwise interactions for the amino acids within a sequence, and so can capture more relevant information from the sequences. Indeed, we show that our adaptation of Transformer yields higher classification accuracy when compared to the recent CNN-based method DeepGO. Second, we modify our model to take motifs in the protein sequences found by BLAST as additional input features. Our strategy is different from other ensemble approaches that average the outcomes of BLAST-based and machine learning predictors. Third, we integrate into our Transformer the metadata about the protein sequences such as 3D structure and protein-protein interaction (PPI) data. We show that such information can greatly improve the prediction accuracy, especially for rare GO labels.
- 2 February 2020
- Published by Cold Spring Harbor Laboratory in bioRxiv
Open AccessAnnotating Gene Ontology terms for protein sequences with the Transformer model
Predicting functions for novel amino acid sequences is a long-standing research problem. The Uniprot database which contains protein sequences annotated with Gene Ontology (GO) terms, is one commonly used training dataset for this problem. Predicting protein functions can then be viewed as a multi-label classification problem where the input is an amino acid sequence and the output is a set of GO terms. Recently, deep convolutional neural network (CNN) models have been introduced to annotate GO terms for protein sequences. However, the CNN architecture can only model close-range interactions between amino acids in a sequence. In this paper, first, we build a novel GO annotation model based on the Transformer neural network. Unlike the CNN architecture, the Transformer models all pairwise interactions for the amino acids within a sequence, and so can capture more relevant information from the sequences. Indeed, we show that our adaptation of Transformer yields higher classification accuracy when compared to the recent CNN-based method DeepGO. Second, we modify our model to take motifs in the protein sequences found by BLAST as additional input features. Our strategy is different from other ensemble approaches that average the outcomes of BLAST-based and machine learning predictors. Third, we integrate into our Transformer the metadata about the protein sequences such as 3D structure and protein-protein interaction (PPI) data. We show that such information can greatly improve the prediction accuracy, especially for rare GO labels.
- Open Access
Evaluating Representations for Gene Ontology Terms
Protein functions can be described by the Gene Ontology (GO) terms, allowing us to compare the functions of two proteins by measuring the similarity of the terms assigned to them. Recent works have applied neural network models to derive the vector representations for GO terms and compute similarity scores for these terms by comparing their vector embeddings. There are two typical ways to embed GO terms into vectors; a model can either embed the definitions of the terms or the topology of the terms in the ontology. In this paper, we design three tasks to critically evaluate the GO embeddings of two recent neural network models, and further introduce additional models for embedding GO terms, adapted from three popular neural network frameworks: Graph Convolution Network (GCN), Embeddings from Language Models (ELMo), and Bidirectional Encoder Representations from Transformers (BERT), which have not yet been explored in previous works. Task 1 studies edge cases where the GO embeddings may not provide meaningful similarity scores for GO terms. We find that all neural network based methods fail to produce high similarity scores for related terms when these terms have low Information Content values. Task 2 is a canonical task which estimates how well GO embeddings can compare functions of two orthologous genes or two interacting proteins. The best neural network methods for this task are those that embed GO terms using their definitions, and the differences among such methods are small. Task 3 evaluates how GO embeddings affect the performance of GO annotation methods, which predict whether a protein should be labeled by certain GO terms. When the annotation datasets contain many samples for each GO label, GO embeddings do not improve the classification accuracy. Machine learning GO annotation methods often remove rare GO labels from the training datasets so that the model parameters can be efficiently trained. We evaluate whether GO embeddings can improve prediction of rare labels unseen in the training datasets, and find that GO embeddings based on the BERT framework achieve the best results in this setting. We present our embedding methods and three evaluation tasks as the basis for future research on this topic.
Open AccessEvaluating Representations for Gene Ontology Terms
Protein functions can be described by the Gene Ontology (GO) terms, allowing us to compare the functions of two proteins by measuring the similarity of the terms assigned to them. Recent works have applied neural network models to derive the vector representations for GO terms and compute similarity scores for these terms by comparing their vector embeddings. There are two typical ways to embed GO terms into vectors; a model can either embed the definitions of the terms or the topology of the terms in the ontology. In this paper, we design three tasks to critically evaluate the GO embeddings of two recent neural network models, and further introduce additional models for embedding GO terms, adapted from three popular neural network frameworks: Graph Convolution Network (GCN), Embeddings from Language Models (ELMo), and Bidirectional Encoder Representations from Transformers (BERT), which have not yet been explored in previous works. Task 1 studies edge cases where the GO embeddings may not provide meaningful similarity scores for GO terms. We find that all neural network based methods fail to produce high similarity scores for related terms when these terms have low Information Content values. Task 2 is a canonical task which estimates how well GO embeddings can compare functions of two orthologous genes or two interacting proteins. The best neural network methods for this task are those that embed GO terms using their definitions, and the differences among such methods are small. Task 3 evaluates how GO embeddings affect the performance of GO annotation methods, which predict whether a protein should be labeled by certain GO terms. When the annotation datasets contain many samples for each GO label, GO embeddings do not improve the classification accuracy. Machine learning GO annotation methods often remove rare GO labels from the training datasets so that the model parameters can be efficiently trained. We evaluate whether GO embeddings can improve prediction of rare labels unseen in the training datasets, and find that GO embeddings based on the BERT framework achieve the best results in this setting. We present our embedding methods and three evaluation tasks as the basis for future research on this topic.
- Open Access
- 20 June 2019
- Published by Public Library of Science (PLoS) in PLOS Biology
- Vol. 17 (6) , e3000333
Challenges and recommendations to improve the installability and archival stability of omics computational tools
Developing new software tools for analysis of large-scale biological data is a key component of advancing modern biomedical research. Scientific reproduction of published findings requires running computational tools on data generated by such studies, yet little attention is presently allocated to the installability and archival stability of computational software tools. Scientific journals require data and code sharing, but none currently require authors to guarantee the continuing functionality of newly published tools. We have estimated the archival stability of computational biology software tools by performing an empirical analysis of the internet presence for 36,702 omics software resources published from 2005 to 2017. We found that almost 28% of all resources are currently not accessible through uniform resource locators (URLs) published in the paper they first appeared in. Among the 98 software tools selected for our installability test, 51% were deemed “easy to install,” and 28% of the tools failed to be installed at all because of problems in the implementation. Moreover, for papers introducing new software, we found that the number of citations significantly increased when authors provided an easy installation process. We propose for incorporation into journal policy several practical solutions for increasing the widespread installability and archival stability of published bioinformatics software.
- 20 June 2019
- Published by Public Library of Science (PLoS) in PLOS Biology
- Vol. 17 (6) , e3000333
Open AccessChallenges and recommendations to improve the installability and archival stability of omics computational tools
Developing new software tools for analysis of large-scale biological data is a key component of advancing modern biomedical research. Scientific reproduction of published findings requires running computational tools on data generated by such studies, yet little attention is presently allocated to the installability and archival stability of computational software tools. Scientific journals require data and code sharing, but none currently require authors to guarantee the continuing functionality of newly published tools. We have estimated the archival stability of computational biology software tools by performing an empirical analysis of the internet presence for 36,702 omics software resources published from 2005 to 2017. We found that almost 28% of all resources are currently not accessible through uniform resource locators (URLs) published in the paper they first appeared in. Among the 98 software tools selected for our installability test, 51% were deemed “easy to install,” and 28% of the tools failed to be installed at all because of problems in the implementation. Moreover, for papers introducing new software, we found that the number of citations significantly increased when authors provided an easy installation process. We propose for incorporation into journal policy several practical solutions for increasing the widespread installability and archival stability of published bioinformatics software.
- Open Access
- 12 July 2017
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 33 (14) , i67-i74
Applying meta-analysis to genotype-tissue expression data from multiple tissues to identify eQTLs and increase the number of eGenes
There is recent interest in using gene expression data to contextualize findings from traditional genome-wide association studies (GWAS). Conditioned on a tissue, expression quantitative trait loci (eQTLs) are genetic variants associated with gene expression, and eGenes are genes whose expression levels are associated with genetic variants. eQTLs and eGenes provide great supporting evidence for GWAS hits and important insights into the regulatory pathways involved in many diseases. When a significant variant or a candidate gene identified by GWAS is also an eQTL or eGene, there is strong evidence to further study this variant or gene. Multi-tissue gene expression datasets like the Gene Tissue Expression (GTEx) data are used to find eQTLs and eGenes. Unfortunately, these datasets often have small sample sizes in some tissues. For this reason, there have been many meta-analysis methods designed to combine gene expression data across many tissues to increase power for finding eQTLs and eGenes. However, these existing techniques are not scalable to datasets containing many tissues, like the GTEx data. Furthermore, these methods ignore a biological insight that the same variant may be associated with the same gene across similar tissues. We introduce a meta-analysis model that addresses these problems in existing methods. We focus on the problem of finding eGenes in gene expression data from many tissues, and show that our model is better than other types of meta-analyses. Source code is at https://github.com/datduong/RECOV. Supplementary data are available at Bioinformatics online.
- 12 July 2017
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 33 (14) , i67-i74
Open AccessApplying meta-analysis to genotype-tissue expression data from multiple tissues to identify eQTLs and increase the number of eGenes
There is recent interest in using gene expression data to contextualize findings from traditional genome-wide association studies (GWAS). Conditioned on a tissue, expression quantitative trait loci (eQTLs) are genetic variants associated with gene expression, and eGenes are genes whose expression levels are associated with genetic variants. eQTLs and eGenes provide great supporting evidence for GWAS hits and important insights into the regulatory pathways involved in many diseases. When a significant variant or a candidate gene identified by GWAS is also an eQTL or eGene, there is strong evidence to further study this variant or gene. Multi-tissue gene expression datasets like the Gene Tissue Expression (GTEx) data are used to find eQTLs and eGenes. Unfortunately, these datasets often have small sample sizes in some tissues. For this reason, there have been many meta-analysis methods designed to combine gene expression data across many tissues to increase power for finding eQTLs and eGenes. However, these existing techniques are not scalable to datasets containing many tissues, like the GTEx data. Furthermore, these methods ignore a biological insight that the same variant may be associated with the same gene across similar tissues. We introduce a meta-analysis model that addresses these problems in existing methods. We focus on the problem of finding eGenes in gene expression data from many tissues, and show that our model is better than other types of meta-analyses. Source code is at https://github.com/datduong/RECOV. Supplementary data are available at Bioinformatics online.
- Open Access
Word and sentence embedding tools to measure semantic similarity of Gene Ontology terms by their definitions
The Gene Ontology (GO) database contains GO terms that describe biological functions of genes. Previous methods for comparing GO terms have relied on the fact that GO terms are organized into a tree structure. Under this paradigm, the locations of two GO terms in the tree dictate their similarity score. In this paper, we introduce two new solutions for this problem, by focusing instead on the definitions of the GO terms. We apply neural network based techniques from the natural language processing (NLP) domain. The first method does not rely on the GO tree, whereas the second indirectly depends on the GO tree. In our first approach, we compare two GO definitions by treating them as two unordered sets of words. The word similarity is estimated by a word embedding model that maps words into an N-dimensional space. In our second approach, we account for the word-ordering within a sentence. We use a sentence encoder to embed GO definitions into vectors and estimate how likely one definition entails another. We validate our methods in two ways. In the first experiment, we test the model’s ability to differentiate a true protein-protein network from a randomly generated network. In the second experiment, we test the model in identifying orthologs from randomly-matched genes in human, mouse, and fly. In both experiments, a hybrid of NLP and GO-tree based method achieves the best classification accuracy.
Open AccessWord and sentence embedding tools to measure semantic similarity of Gene Ontology terms by their definitions
The Gene Ontology (GO) database contains GO terms that describe biological functions of genes. Previous methods for comparing GO terms have relied on the fact that GO terms are organized into a tree structure. Under this paradigm, the locations of two GO terms in the tree dictate their similarity score. In this paper, we introduce two new solutions for this problem, by focusing instead on the definitions of the GO terms. We apply neural network based techniques from the natural language processing (NLP) domain. The first method does not rely on the GO tree, whereas the second indirectly depends on the GO tree. In our first approach, we compare two GO definitions by treating them as two unordered sets of words. The word similarity is estimated by a word embedding model that maps words into an N-dimensional space. In our second approach, we account for the word-ordering within a sentence. We use a sentence encoder to embed GO definitions into vectors and estimate how likely one definition entails another. We validate our methods in two ways. In the first experiment, we test the model’s ability to differentiate a true protein-protein network from a randomly generated network. In the second experiment, we test the model in identifying orthologs from randomly-matched genes in human, mouse, and fly. In both experiments, a hybrid of NLP and GO-tree based method achieves the best classification accuracy.