Using neural networks to mine text and predict metabolic traits for thousands of microbes

Abstract

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.

The phenotype and genotype of fermentative prokaryotes

“…In total, we found that n = 2357 of the 8350 organisms were capable of fermentation or 28% of the total (data file S1). In earlier work with a smaller dataset (19), we found a comparable value (33%).…”

“…Reviews of fermentation have been based on information from model organisms (10)(11)(12)(13)(14)(15), which may not capture the full diversity of this metabolism. Some work, including our own, had started to accumulate information on more prokaryotes (18,19,21,22), but a full picture has still been lacking. By using a dataset of n = 8350 organisms (28% of which are fermentative), the current study paints a fuller picture.…”

“…Reviews of fermentation have been based on information from model organisms ( 10 – 15 ), which may not capture the full diversity of this metabolism. Some work, including our own, had started to accumulate information on more prokaryotes ( 18 , 19 , 21 , 22 ), but a full picture has still been lacking.…”

“…Our laboratory and others have started to collect information on fermentative prokaryotes. Our laboratory has cataloged prokaryotes that carry out fermentation and which form one end product (acetate) ( 18 , 19 ). Another group has cataloged end products of fermentation, but the scope is limited to bacteria of the human gut ( 20 ).…”

The oxidoreductase activity of Rnf balances redox cofactors during fermentation of glucose to propionate in Prevotella

“…Proteomes were searched for proteins using locus tags for genes above. Phylogenetic trees were constructed according to Hackmann and Zhang 80 . We identified habitats of organisms forming propionate, succinate, and acetate using Bergey’s Manual 81 , BacDive 82 , and information from public culture collections.…”