From Genomes to Phenotypes: Traitar, the Microbial Trait Analyzer

Weimann, Aaron; Mooren, Kyra; Frank, Jeremy A.; Pope, Phillip B.; Bremges, Andreas; McHardy, Alice C.

doi:10.1128/msystems.00101-16

Cited by 116 publications

(78 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, a more complete dataset will emerge by integrating information from case studies (e.g., taxa, locations, dates) with information in complementary studies or datasets on species (Weimann et al. ), ecosystems, and ecosystem processes (Yuan et al. ).…”

Section: Discussionmentioning

confidence: 99%

Parasite and pathogen effects on ecosystem processes: A quantitative review

et al. 2020

View full text Add to dashboard Cite

Pathogens and parasites (henceforth “pathogens”) can make up a large percentage of the biomass found in ecosystems, and therefore, their impacts on ecosystem processes should be prominent. Pathogens influence ecosystem processes by affecting the abundance or phenotype of hosts and through direct contributions to ecosystem production. However, there has been little quantitative synthesis of the relative effect sizes of these impacts on ecosystem processes. This study presents a systematic review and meta‐analysis of pathogen effects on primary production, secondary production, and biogeochemical cycles. We find that the effects of pathogens on ecosystem processes were greater where pathogens influenced host or community abundance or biomass than when they influenced phenotypes. Pathogen impacts on primary production were larger than on secondary production or biogeochemical cycles. By contrast, we detected no general differences in effect sizes across host or pathogen taxon or ecosystem type (terrestrial vs. aquatic). While we have found potential evidence of publication bias against negative results, a well‐known issue in meta‐analyses, our work nonetheless shows that the available literature under‐represents some taxa and geographic regions. To better understand the extent and magnitude of pathogen impacts on ecosystem processes, future research is needed in four areas. First, research is needed on the most understudied systems, including bacteria and viruses, as well as tropical ecosystems. A second priority is research seeking to understand how key components of ecosystem variation, including age (time of ecological continuity), productivity, and species diversity and composition, may interact to mediate pathogen impacts. Third, we suggest expanding on work examining how pathogen effects are influenced by climate change, species introductions, deforestation, and other human impacts. Fourth, we expect that host coinfection influences ecosystem processes in ways that cannot always be predicted based on studies of single infections. To enable others to build on this work, we make available the data we extracted from the literature, with the code for computing effect sizes.

show abstract

Section: Discussionmentioning

confidence: 99%

Parasite and pathogen effects on ecosystem processes: A quantitative review

et al. 2020

View full text Add to dashboard Cite

show abstract

“…We encapsulated the sequencing data processing routines in a stand-alone package named seq2geno2pheno https://github.com/hzi-bifo/seq2geno2pheno . The SVM classification was conducted with Model-T https://github.com/hzi-bifo/Model-T , which is built on scikit-learn (Pedregosa et al 2011) and was already used as the prediction engine in our previous work on bacterial trait prediction (Weimann et al 2016) .…”

Section: Methodsmentioning

confidence: 99%

“…We used the support vector machine (SVM) classifier with a linear kernel, as in (Weimann et al 2016) , to predict sensitivity or resistance to four different antibiotics.…”

Section: Figurementioning

confidence: 99%

Fighting antimicrobial resistance in Pseudomonas aeruginosa with machine learning-enabled molecular diagnostics

Khaledi

Weimann

Schniederjans

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

The growing importance of antibiotic resistance on clinical outcomes and cost of care underscores the need for optimization of current diagnostics. For a number of bacterial species antimicrobial resistance can be unambiguously predicted based on their genome sequence. In this study, we sequenced the genomes and transcriptomes of 414 drug-resistant clinical Pseudomonas aeruginosa isolates. By training machine learning classifiers on information about the presence or absence of genes, their sequence variation, and gene expression profiles, we generated predictive models and identified biomarkers of susceptibility or resistance to four commonly administered antimicrobial drugs. Using these data types alone or in combination resulted in high (0.8-0.9) or very high (>0.9) sensitivity and predictive values, where the relative contribution of the different categories of biomarkers strongly depended on the antibiotic. For all drugs except for ciprofloxacin, gene expression information substantially improved diagnostic performance. Our results pave the way for the development of a molecular resistance profiling tool that reliably predicts antimicrobial susceptibility based on genomic and transcriptomic markers.The implementation of a molecular susceptibility test system in routine clinical microbiology diagnostics holds promise to provide earlier and more detailed information on antibiotic resistance profiles of bacterial pathogens and thus could change how physicians treat bacterial infections.

show abstract

“…Zhu et al used a network-based approach to assign functional similarity to pairs of genomes on the basis of encoded proteins 26,27 . Other approaches use direct counts of protein domains to distinguish organisms 28,29 . Both approaches discard context information, which is very important in bacterial and fungal genomes: Not only are genes frequently co-located in e.g.…”

Section: Introductionmentioning

confidence: 99%

An encoding of genome content for machine learning

Viehweger

Krautwurst

Parks

et al. 2019

Preprint

View full text Add to dashboard Cite

14Learning algorithms have at their disposal an ever-growing number of metagenomes for biomining and 15 the study of microbial functions. We propose a novel representation of function called nanotext that 16 scales to very large data sets while capturing precise functional relationships. 17 These relationships are learned from a corpus of 32 thousand genome assemblies with 145 million 18 protein domains. We treat the protein domains in a genome like words in a document, assuming that 19 protein domains in a similar context have similar "meaning". This meaning can be distributed by the 20 Word2Vec embedding algorithm over a vector of numbers. These vectors not only encode function but 21 can be used to predict even complex genomic features and phenotypes. 22 We apply nanotext to data from the Tara ocean expedition to predict plausible culture media and 23 growth temperatures for microorganisms from their metagenome assembled genomes (MAGs) alone. 24 nanotext is freely released under a BSD licence (https://github.com/phiweger/nanotext). 25 An organism can be reduced to the functions its genome encodes. However, the definition of function 27 and its representation remain elusive 1,2 . Protein domains in a genome are basic units of function, like 28 words are basic units of "meaning" in a document. Embeddings of protein domains in a vector space are 29 a novel representation that captures even subtle aspects of function. When extended to entire genomes, 30 functional "topics" of these genomes can be inferred, which reflect their current taxonomy. Domain and 31 genome embeddings have many useful properties especially as input to learning algorithms and offer the 32 possibility for use in large scale metagenomic applications such as biomining and genotype-phenotype 33 mapping. 34In metagenomics, the bottleneck of discovery has shifted from data generation to analysis. Many 35 current sequencing efforts are extremely data-intensive, regularly reconstructing thousands of unknown 36 genomes in a single study [3][4][5][6][7][8] . Gene catalogs compiled from metagenomes have millions to billions 37 of records 9,10 , many without a documented functional role 11 . This wealth of data holds tremendous 38 potential, from substantially revising the tree of life 12 , the discovery of new enzymes and metabolites for 39 biotechnological use 13 to predictive models that distinguish diseases based on microbial composition 14 . 40To adress these questions, learning algorithms such as neural nets are powerful pattern detection 41 tools 15 . Learning is most effective when the signal in the data is "stable", i.e. if given a similar input, 42 the target variable is similar too. Such a stable signal has been found in the functions performed by a 43 microbial community, rather than in its taxonomic makeup 16,17 , although this view is debated 18 . To 44"fit" metagenome-derived functions into learning algorithms, two questions need to be answered: (1) 45 How is "function" defined? (2) How is it represented? 46(1) Pro...

show abstract

From Genomes to Phenotypes: Traitar, the Microbial Trait Analyzer

Cited by 116 publications

References 63 publications

Parasite and pathogen effects on ecosystem processes: A quantitative review

Parasite and pathogen effects on ecosystem processes: A quantitative review

Fighting antimicrobial resistance in Pseudomonas aeruginosa with machine learning-enabled molecular diagnostics

An encoding of genome content for machine learning

Contact Info

Product

Resources

About