A Comprehensive Study on Predicting Functional Role of Metagenomes Using Machine Learning Methods

Abstract

"Metagenomics" is the study of genomic sequences obtained directly from environmental microbial communities with the aim to linking their structures with functional roles. The field has been aided in the unprecedented advancement through high-throughput omics data sequencing. The outcome of sequencing are biologically rich data sets. Metagenomic data consisting of microbial spe-cies which outnumber microbial samples, lead to the "curse of dimensionality". Hence the focus in metagenomics studies has moved towards developing efficient computational models using Machine Learning (ML), reducing the computational cost. In this paper, we comprehensively assessed various ML approaches to classifying high-dimensional human microbiota effectively into their functional phenotypes. We propose the application of embedded feature selection methods, namely, Extreme Gradient Boost-ing and Penalized Logistic Regression to determine important species. The resultant feature set enhanced the performance of one of the most popular state-of-the-art methods, Random Forest (RF) over metagenomic studies. Experimental results indicate that the proposed method achieved best results in terms of accuracy, area under Receiver Operating Characteristic curve (ROC-AUC) and major improvement in processing time. It outperformed other feature selection methods of filters or wrappers over RF and classifiers such as Support Vector Machine (SVM), Extreme Learning Machine (ELM), and -Nearest Neighbors (-NN).

An Integrative Framework for Functional Analysis of Cattle Rumen Microbiomes

“…The model outputs the predictive class majority voting amongst the constituent individual trees [21]. Several studies have applied RF for the prediction of metagenomic functions [7][8][9]. b. Reg-LR.…”

“…We found that the incorporation of the phylogenetic tree structure into analysis has potential to increase the prediction power. More accurate predictions were obtained with the phylogeny-aware pipeline as proposed in [11], when benchmarked with other nonphylogenetic [7][8][9] and phylogenetic-based approaches [12][13][14]. Some high-level visualizations and their applicability to our data are also presented.…”

“…RF has been established as a gold-standard method for classifying the metagenomes [7,8]. Wassan et al [9], suggested regularized-LR (reg-LR) and extreme gradient Boosting (XgBoost) ensemble with RF as a potential method in metagenomic data. Pasolli et al [15] suggested feature engineering by RF is effective for functional meta-analysis of OTU abundances.…”

“…This resulted in an integrative ML pipeline to investigate the presence of related species (i.e. OTUs/taxas/ASVs) and their role in determining phenotype unlike studies which considered the microbial species as independent features [7][8][9][10]. We found that the incorporation of the phylogenetic tree structure into analysis has potential to increase the prediction power.…”

Đánh Giá Các Công Cụ Hỗ Trợ Chẩn Đoán Bệnh Với Cách Tiếp Cận Y Học Cá Thể Hóa Trên Dữ Liệu Metagenomic

“…Các thí nghiệm đã được tiến hành để dự đoán bệnh dựa trên dữ liệu Metagenomic, sử dụng nền tảng trong R-project ở trang web http://www.R-project.org, phiên bản 3.0.4. (Wassan, 2018). Các gói khác nhau liên quan đến các mô hình máy học-Machine Learning (ML) và bộ thông số cấu hình tối ưu có liên quan được sử dụng trong nghiên cứu này được liệt kê dưới đây (Bảng 2).…”