Microbes can metabolize more chemical compounds than any other group of organisms. As a result, their metabolism is of interest to investigators across biology. Despite the interest, information on metabolism of specific microbes is hard to access. Information is buried in text of books and journals, and investigators have no easy way to extract it out. Here we investigate if neural networks can extract out this information and predict metabolic traits. For proof of concept, we predicted two traits: whether microbes carry one type of metabolism (fermentation) or produce one metabolite (acetate). We collected written descriptions of 7,021 species of bacteria and archaea from Bergey's Manual. We read the descriptions and manually identified (labeled) which species were fermentative or produced acetate. We then trained neural networks to predict these labels. In total, we identified 2,364 species as fermentative, and 1,009 species as also producing acetate. Neural networks could predict which species were fermentative with 97.3% accuracy. Accuracy was even higher (98.6%) when predicting species also producing acetate. Phylogenetic trees of species and their traits confirmed that predictions were accurate. Our approach with neural networks can extract information efficiently and accurately. It paves the way for putting more metabolic traits into databases, providing easy access of information to investigators.

MOTIVATION: Microbes are the most diverse organisms on the planet. Deep sequencing of ribosomal DNA (rDNA) suggests thousands of different microbes may be present in a single sample. However, errors in sequencing have made any estimate of within-sample (alpha) diversity uncertain. RESULTS: We developed a tool to estimate alpha diversity of rDNA sequences from microbes (and other sequences). Our tool, Distanced, calculates how different (distant) sequences would be without sequencing errors. It does this using a Bayesian approach. Using this approach, Distanced accurately estimated alpha diversity of rDNA sequences from bacteria and fungi. It had lower root mean square prediction error (RMSPE) than when using no tool (leaving sequencing errors uncorrected). It was also accurate with non-microbial sequences (antibody mRNA). State-of-the-art tools (DADA2 and Deblur) were far less accurate. They often had higher RMSPE than when using no tool. Distanced thus represents an improvement over existing tools. Distanced will be useful to several disciplines, given microbial diversity affects everything from human health to ecosystem function. AVAILABILITY AND IMPLEMENTATION: Distanced is freely available at https://github.com/thackmann/Distanced. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.