Systematic evaluation of supervised machine learning for sample origin prediction using metagenomic sequencing data

Abstract

Background:The advent of metagenomic sequencing provides microbial abundance patterns that can be leveraged for sample origin prediction. Supervised machine learning classification approaches have been reported to predict sample origin accurately when the origin has been previously sampled. Using metagenomic datasets provided by the 2019 CAMDA challenge, we evaluated the influence of technical, analytical and machine learning approaches for result interpretation and source prediction of new origins.Results:Comparison between 16S rRNA amplicon and shotgun sequencing approaches as well as metagenomic analytical tools showed differences in measured microbial abundance of the same samples, especially for organisms present at low abundance. Shotgun sequence data analyzed using Kraken2 and Bracken taxonomic annotation, had higher detection sensitivity than did other methods. As classification models are limited to labeling previously trained origins, we proposed an alternative approach using Lasso-regularized multivariate regression to predict geographic coordinates for comparison. In both models, the prediction errors were much higher in Leave-1-city-out than in 10-fold cross validation, the former of which realistically forecasted the difficulty in accurately predicting samples from new origins than pre-trained origins. The challenge was further confirmed using mystery samples obtained from new origins. Overall, prediction performances between regression and classification models, as measured by mean squared error, were comparable on mystery samples. Due to higher prediction errors for samples from new origins, we provided an additional strategy based on prediction ambiguity to infer whether a sample is from a new origin for practical applications. Lastly, we showed increased prediction error when data from a different sequencing protocol were included as training data.Conclusions:Here we highlighted the capacity of predicting sample origin accurately with pre-trained origins and the challenge of predicting new origins through both regression and classification models. Overall, the work provided a summary evaluation of sequencing techniques, protocol, taxonomic analytical approaches, and machine learning approaches to inform future designs in metagenomic prediction of sample origin.

NUAK2 and RCan2 participate in the p53 mutant pro-tumorigenic network

“…Cancer genomic studies have seen an exponential growth in the last decade [57][58][59][60], and a massive amount of data has now become openly accessible. Analyses of this information can effectively direct studies addressing the activation of specific gene network and their prognostic significance in cancer pathogenesis [61][62][63].…”

Global mapping of cancers: The Cancer Genome Atlas and beyond

“…Artificial intelligence might just help with this issue. The possibility to develop executable cancer models would allow scientists to search among multiple datasets for the discovery of signatures at every level (genomic, transcriptomic, proteomic, clinical data) to detect key features of the biological behaviour of interest [20,69,[165][166][167]. Biological systems can be treated by AI as networks of information that can be programmed, i.e., reconstructed as a matrix of data [168,169].…”

Serine and one-carbon metabolisms bring new therapeutic venues in prostate cancer

“…In the pathogenesis of prostate cancer, the initial pathogenetic event might be considered the dependency from the androgen stimulation [46][47][48][49][50], but many more molecular alterations have been found associated to its progression [51][52][53][54][55] as well as to the development of specific clinical features of the disease such as bone metastasis [56] or studied as potential therapeutic targets [57]. The microbiome has also been shown to impact on the pathogenesis of this kind of cancer [58][59][60][61][62], also due to development of algorithms allowing to interpret its role in the context of a disease [63][64][65][66][67]. Anyhow, specific attention is focused on epigenetic remodeling.…”