Beating Naive Bayes at Taxonomic Classification of 16S rRNA Gene Sequences

Ziemski, Michał; Wisanwanichthan, Treepop; Bokulich, Nicholas A.; Kaehler, Benjamin

doi:10.3389/fmicb.2021.644487

Cited by 19 publications

(19 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, the naive bayes requires more sample to train compared to the neural network in this work may be because of the data used. Similar results that Naive Bayes classifier outperformed the neural network classifier have been reported in different applications [ 48 ] (see [ 49 ] for a review). Additionally, sample size significantly influences the performance of classifiers, as seen in our results and in a review study that the large sample sizes depict relatively precise and similar accuracies among classifiers [ 47 ].…”

Section: Discussionsupporting

confidence: 82%

Evaluation of a decided sample size in machine learning applications

2023

View full text Add to dashboard Cite

Background An appropriate sample size is essential for obtaining a precise and reliable outcome of a study. In machine learning (ML), studies with inadequate samples suffer from overfitting of data and have a lower probability of producing true effects, while the increment in sample size increases the accuracy of prediction but may not cause a significant change after a certain sample size. Existing statistical approaches using standardized mean difference, effect size, and statistical power for determining sample size are potentially biased due to miscalculations or lack of experimental details. This study aims to design criteria for evaluating sample size in ML studies. We examined the average and grand effect sizes and the performance of five ML methods using simulated datasets and three real datasets to derive the criteria for sample size. We systematically increase the sample size, starting from 16, by randomly sampling and examine the impact of sample size on classifiers’ performance and both effect sizes. Tenfold cross-validation was used to quantify the accuracy. Results The results demonstrate that the effect sizes and the classification accuracies increase while the variances in effect sizes shrink with the increment of samples when the datasets have a good discriminative power between two classes. By contrast, indeterminate datasets had poor effect sizes and classification accuracies, which did not improve by increasing sample size in both simulated and real datasets. A good dataset exhibited a significant difference in average and grand effect sizes. We derived two criteria based on the above findings to assess a decided sample size by combining the effect size and the ML accuracy. The sample size is considered suitable when it has appropriate effect sizes (≥ 0.5) and ML accuracy (≥ 80%). After an appropriate sample size, the increment in samples will not benefit as it will not significantly change the effect size and accuracy, thereby resulting in a good cost-benefit ratio. Conclusion We believe that these practical criteria can be used as a reference for both the authors and editors to evaluate whether the selected sample size is adequate for a study.

show abstract

Section: Discussionsupporting

confidence: 82%

Evaluation of a decided sample size in machine learning applications

2023

View full text Add to dashboard Cite

show abstract

“…The Quantitative Insights into Microbial Ecology 2 (QIIME 2 version 2021.8) [20] was used to quality lter, denoise and analyze the sequences. Demultiplexed paired-end sequence reads were denoised with DADA2 into amplicon sequence variants (ASV).…”

Section: Discussionmentioning

confidence: 99%

Gut Microbiome Variation in Pulmonary TB Patients with Diabetes or HIV Comorbidities

Morgan

Parbie

Boadu

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: The gut microbiota is known to play critical role in shaping the host immunity, and metabolism and influences the onset and progression of both communicable and non-communicable diseases. This study assessed the gut microbiome of tuberculosis (TB) cases with diabetes mellitus (DM) or HIV comorbidities before anti-TB therapy and after the intensive phase anti-TB therapy. Methods: Ninety cases comprising 60 TB-only, 23 TB-DM, 7 TB-HIV were recruited, among which 35 TB-only, 10 TB-DM, 5 TB-HIV were also sampled after 2 months of anti-TB treatment. Total gut microbiome was detected by 16S rRNA gene sequencing of DNA extracted from collected stool specimen. The taxonomic and functional diversity of the different groups were compared in addition to changes that could occur after 2 months antibiotics use. Results: Compared to the healthy controls, the gut microbiome of all the TB cohorts was characterized by a significant decreased alpha diversity and significant compositional changes. All the three TB cohorts were enriched with inflammatory related microorganisms of the genera Escherichia-shigella, Streptococcus, Enterococcus and Erysipelatoclostridium with depletion in beneficial taxa of the genera Faecalibacterium, Bifidobacterium and Clostridium. In pairwise comparison with the healthy controls, the TB-only cohort were enriched with Streptococcus and Erysipelatoclostridium, the TB-DM enriched with Bacteroides, and TB-HIV enriched with Escherichia-shigella, Dialister and Erysipelatoclostridium. After the intensive phase anti-TB therapy, there was general enrichment of the genera Erysipelotrichaceae_UCG 003, Veillonella and Fusobacterium.Conclusion: Our findings show a dysbiotic gut microbiome and associated increase in the number of inflammation related microorganism in gut of TB individuals with or without comorbidity.

show abstract

“…Feature table and feature data summaries were generated to determine the sequence distribution per sample as well as a-diversity (abundance of ASV within a sample) and bdiversity (microbial composition between samples) analysis of the samples. Taxonomic classi cations were performed using Naïve Bayesian classi er [21] against the SILVA database release 138 [22] trimmed to the V3-V4 region of the 16S rRNA gene.…”

Section: Discussionmentioning

confidence: 99%

Gut Microbiome Variation in Pulmonary TB Patients with Diabetes or HIV Comorbidities

Morgan

Parbie

Boadu

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: The gut microbiota is known to play critical role in shaping the host immunity, and metabolism and influences the onset and progression of both communicable and non-communicable diseases. This study assessed the gut microbiome of tuberculosis (TB) cases with diabetes mellitus (DM) or HIV comorbidities before anti-TB therapy and after the intensive phase anti-TB therapy. Methods: Ninety cases comprising 60 TB-only, 23 TB-DM, 7 TB-HIV were recruited, among which 35 TB-only, 10 TB-DM, 5 TB-HIV were also sampled after 2 months of anti-TB treatment. Total gut microbiome was detected by 16S rRNA gene sequencing of DNA extracted from collected stool specimen. The taxonomic and functional diversity of the different groups were compared in addition to changes that could occur after 2 months antibiotics use. Results: Compared to the healthy controls, the gut microbiome of all the TB cohorts was characterized by a significant decreased alpha diversity and significant compositional changes. All the three TB cohorts were enriched with inflammatory related microorganisms of the genera Escherichia-shigella, Streptococcus, Enterococcus and Erysipelatoclostridiumwith depletion in beneficial taxa of the genera Faecalibacterium, Bifidobacteriumand Clostridium. In pairwise comparison with the healthy controls, the TB-only cohort were enriched with Streptococcus and Erysipelatoclostridium, the TB-DM enriched with Bacteroides, and TB-HIV enriched with Escherichia-shigella, Dialister and Erysipelatoclostridium. After the intensive phase anti-TB therapy, there was general enrichment of the genera Erysipelotrichaceae_UCG 003, Veillonella and Fusobacterium. Conclusion: Our findings show a dysbiotic gut microbiome and associated increase in the number of inflammation related microorganism in gut of TB individuals with or without comorbidity. Funding: European and Developing Countries Clinical Trials Partnership (EDCTP) Senior Fellowship to Dorothy Yeboah-Manu (TMA-2017-GSF-1942)

show abstract

Beating Naive Bayes at Taxonomic Classification of 16S rRNA Gene Sequences

Cited by 19 publications

References 31 publications

Evaluation of a decided sample size in machine learning applications

Evaluation of a decided sample size in machine learning applications

Gut Microbiome Variation in Pulmonary TB Patients with Diabetes or HIV Comorbidities

Gut Microbiome Variation in Pulmonary TB Patients with Diabetes or HIV Comorbidities

Contact Info

Product

Resources

About