@article{
 title = {Dynamic simulations of microbial communities under perturbations: opportunities for microbiome engineering},
 type = {article},
 year = {2020},
 websites = {https://www.researchsquare.com/article/rs-14990/v1},
 month = {2},
 day = {25},
 id = {50bc36be-bacc-327e-936b-0e499716f5e8},
 created = {2020-08-12T08:24:06.671Z},
 accessed = {2020-07-30},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2020-08-12T08:24:06.671Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 citation_key = {Garcia-Jimenez2020},
 private_publication = {false},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and Carrasco, Jorge and Medina, Joaquín and Wilkinson, Mark D.},
 doi = {10.21203/RS.2.24431/V1},
 journal = {Research Square - Preprint},
 number = {Version 1}
}

@article{
 title = {Data-driven classification of the certainty of scholarly assertions},
 type = {article},
 year = {2020},
 keywords = {Certainty,FAIR Data,Machine learning,Scholarly communication,Text mining},
 volume = {2020},
 id = {4cc2767e-8fd1-301e-8238-77fe826b1837},
 created = {2020-08-12T08:24:06.673Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2020-08-12T08:24:06.673Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {© Copyright 2020 Prieto et al. The grammatical structures scholars use to express their assertions are intended to convey various degrees of certainty or speculation. Prior studies have suggested a variety of categorization systems for scholarly certainty; however, these have not been objectively tested for their validity, particularly with respect to representing the interpretation by the reader, rather than the intention of the author. In this study, we use a series of questionnaires to determine how researchers classify various scholarly assertions, using three distinct certainty classification systems. We find that there are three distinct categories of certainty along a spectrum from high to low. We show that these categories can be detected in an automated manner, using a machine learning model, with a cross-validation accuracy of 89.2% relative to an author-annotated corpus, and 82.2% accuracy against a publicly-annotated corpus. This finding provides an opportunity for contextual metadata related to certainty to be captured as a part of text-mining pipelines, which currently miss these subtle linguistic cues. We provide an exemplar machine-accessible representation-a Nanopublication-where certainty category is embedded as metadata in a formal, ontology-based manner within text-mined scholarly assertions.},
 bibtype = {article},
 author = {Prieto, M. and Deus, H. and de Waard, A. and Schultes, E. and García-Jiménez, B. and Wilkinson, M.D.},
 doi = {10.7717/peerj.8871},
 journal = {PeerJ},
 number = {4}
}

© Copyright 2020 Prieto et al. The grammatical structures scholars use to express their assertions are intended to convey various degrees of certainty or speculation. Prior studies have suggested a variety of categorization systems for scholarly certainty; however, these have not been objectively tested for their validity, particularly with respect to representing the interpretation by the reader, rather than the intention of the author. In this study, we use a series of questionnaires to determine how researchers classify various scholarly assertions, using three distinct certainty classification systems. We find that there are three distinct categories of certainty along a spectrum from high to low. We show that these categories can be detected in an automated manner, using a machine learning model, with a cross-validation accuracy of 89.2% relative to an author-annotated corpus, and 82.2% accuracy against a publicly-annotated corpus. This finding provides an opportunity for contextual metadata related to certainty to be captured as a part of text-mining pipelines, which currently miss these subtle linguistic cues. We provide an exemplar machine-accessible representation-a Nanopublication-where certainty category is embedded as metadata in a formal, ontology-based manner within text-mined scholarly assertions.

@article{
 title = {Predicting microbiomes through a deep latent space},
 type = {article},
 year = {2020},
 pages = {2020.04.27.063974},
 websites = {http://biorxiv.org/content/early/2020/04/28/2020.04.27.063974.abstract},
 day = {28},
 id = {1afe7444-ae88-366b-adfe-dd830cad559e},
 created = {2020-08-12T08:24:06.771Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2020-08-12T08:24:06.771Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 source_type = {Preprint},
 private_publication = {false},
 abstract = {Motivation Microbial communities influence their environment by modifying the availability of compounds such as nutrients or chemical elicitors. Knowing the microbial composition of a site is therefore relevant to improving productivity or health. However, sequencing facilities are not always available, or may be prohibitively expensive in some cases. Thus, it would be desirable to computationally predict the microbial composition from more accessible, easily-measured features.Results Integrating Deep Learning techniques with microbiome data, we propose an artificial neural network architecture based on heterogeneous autoencoders to condense the long vector of microbial abundance values into a deep latent space representation. Then, we design a model to predict the deep latent space and, consequently, to predict the complete microbial composition using environmental features as input. The performance of our system is examined using the rhizosphere microbiome of Maize. We reconstruct the microbial composition (717 taxa) from the deep latent space (10 values) with high fidelity (¿0.9 Pearson correlation). We then successfully predict microbial composition from environmental variables such as plant age, temperature or precipitation (0.73 Pearson correlation, 0.42 Bray-Curtis). We extend this to predict microbiome composition under hypothetical scenarios, such as future climate change conditions. Finally, via transfer learning, we predict microbial composition in a distinct scenario with only a hundred sequences, and distinct environmental features. We propose that our deep latent space may assist microbiome-engineering strategies when technical or financial resources are limited, through predicting current or future microbiome compositions.Availability Software, results, and data are available at https://github.com/jorgemf/DeepLatentMicrobiomeCompeting Interest StatementThe authors have declared no competing interest.},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and Muñoz, Jorge and Cabello, Sara and Medina, Joaquín and Wilkinson, Mark D},
 doi = {10.1101/2020.04.27.063974},
 journal = {bioRxiv}
}

Motivation Microbial communities influence their environment by modifying the availability of compounds such as nutrients or chemical elicitors. Knowing the microbial composition of a site is therefore relevant to improving productivity or health. However, sequencing facilities are not always available, or may be prohibitively expensive in some cases. Thus, it would be desirable to computationally predict the microbial composition from more accessible, easily-measured features.Results Integrating Deep Learning techniques with microbiome data, we propose an artificial neural network architecture based on heterogeneous autoencoders to condense the long vector of microbial abundance values into a deep latent space representation. Then, we design a model to predict the deep latent space and, consequently, to predict the complete microbial composition using environmental features as input. The performance of our system is examined using the rhizosphere microbiome of Maize. We reconstruct the microbial composition (717 taxa) from the deep latent space (10 values) with high fidelity (¿0.9 Pearson correlation). We then successfully predict microbial composition from environmental variables such as plant age, temperature or precipitation (0.73 Pearson correlation, 0.42 Bray-Curtis). We extend this to predict microbiome composition under hypothetical scenarios, such as future climate change conditions. Finally, via transfer learning, we predict microbial composition in a distinct scenario with only a hundred sequences, and distinct environmental features. We propose that our deep latent space may assist microbiome-engineering strategies when technical or financial resources are limited, through predicting current or future microbiome compositions.Availability Software, results, and data are available at https://github.com/jorgemf/DeepLatentMicrobiomeCompeting Interest StatementThe authors have declared no competing interest.

@article{
 title = {MEMOTE for standardized genome-scale metabolic model testing},
 type = {article},
 year = {2020},
 volume = {38},
 id = {0357e95d-6708-39a2-9d78-16126c8ad6ab},
 created = {2020-08-12T08:24:06.806Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2020-08-12T08:24:06.806Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 citation_key = {Lieven2020},
 private_publication = {false},
 bibtype = {article},
 author = {Lieven, C. and Beber, M.E. and Olivier, B.G. and Bergmann, F.T. and Ataman, M. and Babaei, P. and Bartell, J.A. and Blank, L.M. and Chauhan, S. and Correia, K. and Diener, C. and Dräger, A. and Ebert, B.E. and Edirisinghe, J.N. and Faria, J.P. and Feist, A.M. and Fengos, G. and Fleming, R.M.T. and García-Jiménez, B. and Hatzimanikatis, V. and van Helvoirt, W. and Henry, C.S. and Hermjakob, H. and Herrgård, M.J. and Kaafarani, A. and Kim, H.U. and King, Z. and Klamt, S. and Klipp, E. and Koehorst, J.J. and König, M. and Lakshmanan, M. and Lee, D.-Y. and Lee, S.Y. and Lee, S. and Lewis, N.E. and Liu, F. and Ma, H. and Machado, D. and Mahadevan, R. and Maia, P. and Mardinoglu, A. and Medlock, G.L. and Monk, J.M. and Nielsen, J. and Nielsen, L.K. and Nogales, J. and Nookaew, I. and Palsson, B.O. and Papin, J.A. and Patil, K.R. and Poolman, M. and Price, N.D. and Resendis-Antonio, O. and Richelle, A. and Rocha, I. and Sánchez, B.J. and Schaap, P.J. and Malik Sheriff, R.S. and Shoaie, S. and Sonnenschein, N. and Teusink, B. and Vilaça, P. and Vik, J.O. and Wodke, J.A.H. and Xavier, J.C. and Yuan, Q. and Zakhartsev, M. and Zhang, C.},
 doi = {10.1038/s41587-020-0446-y},
 journal = {Nature Biotechnology},
 number = {3}
}

@article{
 title = {Metabolic Modelling Approaches for Describing and Engineering Microbial Communities},
 type = {article},
 year = {2020},
 keywords = {computational methods,design,engineering,genome-scale metabolic model,microbial community,optimization,synthetic microbial consortia},
 pages = {1-53},
 websites = {https://www.preprints.org/manuscript/202009.0548/v1,https://authors.elsevier.com/tracking/article/details.do?surname=Nogales&aid=777&jid=CSBJ},
 publisher = {Computational and Structural Biotechnology Journal},
 day = {5},
 id = {3309cc4b-efb1-3588-83a5-44e22cb6889c},
 created = {2020-09-24T08:15:35.890Z},
 accessed = {2020-12-05},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2020-12-23T17:05:25.471Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and Torres, Jesús and Nogales, Juan},
 doi = {10.20944/PREPRINTS202009.0548.V1},
 journal = {Preprints.org},
 number = {December}
}

@article{
 title = {Predicting microbiomes through a deep latent space},
 type = {article},
 year = {2020},
 websites = {https://academic.oup.com/bioinformatics/advance-article/doi/10.1093/bioinformatics/btaa971/5988714},
 month = {11},
 publisher = {Oxford University Press (OUP)},
 day = {18},
 id = {f09b729e-ce43-3683-ada2-52abf18a9d79},
 created = {2020-12-23T17:02:17.349Z},
 accessed = {2020-12-14},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2020-12-23T17:02:17.349Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {Motivation Microbial communities influence their environment by modifying the availability of compounds such as nutrients or chemical elicitors. Knowing the microbial composition of a site is therefore relevant to improving productivity or health. However, sequencing facilities are not always available, or may be prohibitively expensive in some cases. Thus, it would be desirable to computationally predict the microbial composition from more accessible, easily-measured features.Results Integrating Deep Learning techniques with microbiome data, we propose an artificial neural network architecture based on heterogeneous autoencoders to condense the long vector of microbial abundance values into a deep latent space representation. Then, we design a model to predict the deep latent space and, consequently, to predict the complete microbial composition using environmental features as input. The performance of our system is examined using the rhizosphere microbiome of Maize. We reconstruct the microbial composition (717 taxa) from the deep latent space (10 values) with high fidelity (¿0.9 Pearson correlation). We then successfully predict microbial composition from environmental variables such as plant age, temperature or precipitation (0.73 Pearson correlation, 0.42 Bray-Curtis). We extend this to predict microbiome composition under hypothetical scenarios, such as future climate change conditions. Finally, via transfer learning, we predict microbial composition in a distinct scenario with only a hundred sequences, and distinct environmental features. We propose that our deep latent space may assist microbiome-engineering strategies when technical or financial resources are limited, through predicting current or future microbiome compositions.Availability Software, results, and data are available at https://github.com/jorgemf/DeepLatentMicrobiomeCompeting Interest StatementThe authors have declared no competing interest.},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and Muñoz, Jorge and Cabello, Sara and Medina, Joaquín and Wilkinson, Mark D},
 editor = {Jonathan, Wren},
 doi = {10.1093/bioinformatics/btaa971},
 journal = {Bioinformatics}
}

Motivation Microbial communities influence their environment by modifying the availability of compounds such as nutrients or chemical elicitors. Knowing the microbial composition of a site is therefore relevant to improving productivity or health. However, sequencing facilities are not always available, or may be prohibitively expensive in some cases. Thus, it would be desirable to computationally predict the microbial composition from more accessible, easily-measured features.Results Integrating Deep Learning techniques with microbiome data, we propose an artificial neural network architecture based on heterogeneous autoencoders to condense the long vector of microbial abundance values into a deep latent space representation. Then, we design a model to predict the deep latent space and, consequently, to predict the complete microbial composition using environmental features as input. The performance of our system is examined using the rhizosphere microbiome of Maize. We reconstruct the microbial composition (717 taxa) from the deep latent space (10 values) with high fidelity (¿0.9 Pearson correlation). We then successfully predict microbial composition from environmental variables such as plant age, temperature or precipitation (0.73 Pearson correlation, 0.42 Bray-Curtis). We extend this to predict microbiome composition under hypothetical scenarios, such as future climate change conditions. Finally, via transfer learning, we predict microbial composition in a distinct scenario with only a hundred sequences, and distinct environmental features. We propose that our deep latent space may assist microbiome-engineering strategies when technical or financial resources are limited, through predicting current or future microbiome compositions.Availability Software, results, and data are available at https://github.com/jorgemf/DeepLatentMicrobiomeCompeting Interest StatementThe authors have declared no competing interest.

@article{
 title = {Metabolic Modelling Approaches for Describing and Engineering Microbial Communities},
 type = {article},
 year = {2020},
 keywords = {computational methods,design,engineering,genome-scale metabolic model,microbial community,optimization,synthetic microbial consortia},
 pages = {1-53},
 websites = {https://doi.org/10.1016/j.csbj.2020.12.003},
 publisher = {Elsevier},
 day = {5},
 id = {cb87e632-47cd-3693-b6d0-68dfbe8ea10c},
 created = {2020-12-23T17:02:17.972Z},
 accessed = {2020-12-05},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2020-12-23T17:02:17.972Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and Torres, Jesús and Nogales, Juan},
 doi = {10.1016/j.csbj.2020.12.003},
 journal = {Computational and Structural Biotechnology Journal},
 number = {December}
}

@article{
 title = {Publisher Correction: MEMOTE for standardized genome-scale metabolic model testing (Nature Biotechnology, (2020), 38, 3, (272-276), 10.1038/s41587-020-0446-y)},
 type = {article},
 year = {2020},
 volume = {38},
 id = {c94e7f63-1b76-39c1-b38b-cabbd3aff1aa},
 created = {2020-12-23T17:02:18.149Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2020-12-23T17:02:18.149Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {© 2020, The Author(s). An amendment to this paper has been published and can be accessed via a link at the top of the paper.},
 bibtype = {article},
 author = {Lieven, C. and Beber, M.E. and Olivier, B.G. and Bergmann, F.T. and Ataman, M. and Babaei, P. and Bartell, J.A. and Blank, L.M. and Chauhan, S. and Correia, K. and Diener, C. and Dräger, A. and Ebert, B.E. and Edirisinghe, J.N. and Faria, J.P. and Feist, A.M. and Fengos, G. and Fleming, R.M.T. and García-Jiménez, B. and Hatzimanikatis, V. and van Helvoirt, W. and Henry, C.S. and Hermjakob, H. and Herrgård, M.J. and Kaafarani, A. and Kim, H.U. and King, Z. and Klamt, S. and Klipp, E. and Koehorst, J.J. and König, M. and Lakshmanan, M. and Lee, D.-Y. and Lee, S.Y. and Lee, S. and Lewis, N.E. and Liu, F. and Ma, H. and Machado, D. and Mahadevan, R. and Maia, P. and Mardinoglu, A. and Medlock, G.L. and Monk, J.M. and Nielsen, J. and Nielsen, L.K. and Nogales, J. and Nookaew, I. and Palsson, B.O. and Papin, J.A. and Patil, K.R. and Poolman, M. and Price, N.D. and Resendis-Antonio, O. and Richelle, A. and Rocha, I. and Sánchez, B.J. and Schaap, P.J. and Sheriff, R.S.M. and Shoaie, S. and Sonnenschein, N. and Teusink, B. and Vilaça, P. and Vik, J.O. and Wodke, J.A.H. and Xavier, J.C. and Yuan, Q. and Zakhartsev, M. and Zhang, C.},
 doi = {10.1038/s41587-020-0477-4},
 journal = {Nature Biotechnology},
 number = {4}
}

© 2020, The Author(s). An amendment to this paper has been published and can be accessed via a link at the top of the paper.

@article{
 title = {Robust and automatic definition of microbiome states},
 type = {article},
 year = {2019},
 keywords = {Clustering,Longitudinal dataset,Machine Learning,Metagenomics,Microbiome,Sub-states},
 pages = {e6657},
 volume = {7},
 websites = {https://peerj.com/articles/6657},
 month = {3},
 publisher = {PeerJ Inc.},
 day = {26},
 id = {41623699-88cb-3079-9ba3-8f1bf45a665f},
 created = {2019-09-30T15:41:16.319Z},
 accessed = {2019-09-30},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2019-09-30T17:15:18.279Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {Analysis of microbiome dynamics would allow elucidation of patterns within microbial community evolution under a variety of biologically or economically important circumstances; however, this is currently hampered in part by the lack of rigorous, formal, yet generally-applicable approaches to discerning distinct configurations of complex microbial populations. Clustering approaches to define microbiome “community state-types” at a population-scale are widely used, though not yet standardized. Similarly, distinct variations within a state-type are well documented, but there is no rigorous approach to discriminating these more subtle variations in community structure. Finally, intra-individual variations with even fewer differences will likely be found in, for example, longitudinal data, and will correlate with important features such as sickness versus health. We propose an automated, generic, objective, domain-independent, and internally-validating procedure to define statistically distinct microbiome states within datasets containing any degree of phylotypic diversity. Robustness of state identification is objectively established by a combination of diverse techniques for stable cluster verification. To demonstrate the efficacy of our approach in detecting discreet states even in datasets containing highly similar bacterial communities, and to demonstrate the broad applicability of our method, we reuse eight distinct longitudinal microbiome datasets from a variety of ecological niches and species. We also demonstrate our algorithm’s flexibility by providing it distinct taxa subsets as clustering input, demonstrating that it operates on filtered or unfiltered data, and at a range of different taxonomic levels. The final output is a set of robustly defined states which can then be used as general biomarkers for a wide variety of downstream purposes such as association with disease, monitoring response to intervention, or identifying optimally performant populations.},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and Wilkinson, Mark D.},
 doi = {10.7717/peerj.6657},
 journal = {PeerJ}
}

Analysis of microbiome dynamics would allow elucidation of patterns within microbial community evolution under a variety of biologically or economically important circumstances; however, this is currently hampered in part by the lack of rigorous, formal, yet generally-applicable approaches to discerning distinct configurations of complex microbial populations. Clustering approaches to define microbiome “community state-types” at a population-scale are widely used, though not yet standardized. Similarly, distinct variations within a state-type are well documented, but there is no rigorous approach to discriminating these more subtle variations in community structure. Finally, intra-individual variations with even fewer differences will likely be found in, for example, longitudinal data, and will correlate with important features such as sickness versus health. We propose an automated, generic, objective, domain-independent, and internally-validating procedure to define statistically distinct microbiome states within datasets containing any degree of phylotypic diversity. Robustness of state identification is objectively established by a combination of diverse techniques for stable cluster verification. To demonstrate the efficacy of our approach in detecting discreet states even in datasets containing highly similar bacterial communities, and to demonstrate the broad applicability of our method, we reuse eight distinct longitudinal microbiome datasets from a variety of ecological niches and species. We also demonstrate our algorithm’s flexibility by providing it distinct taxa subsets as clustering input, demonstrating that it operates on filtered or unfiltered data, and at a range of different taxonomic levels. The final output is a set of robustly defined states which can then be used as general biomarkers for a wide variety of downstream purposes such as association with disease, monitoring response to intervention, or identifying optimally performant populations.

@article{
 title = {Data-driven classification of the certainty of scholarly assertions},
 type = {article},
 year = {2019},
 keywords = {FAIR Data,certainty,machine learning,scholarly communication,text mining},
 websites = {https://peerj.com/preprints/27829/},
 month = {6},
 publisher = {PeerJ Inc.},
 day = {27},
 id = {f8f29c4e-8289-39ad-910f-7e8501bc3c10},
 created = {2019-09-30T15:42:19.863Z},
 accessed = {2019-09-30},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2019-09-30T15:42:19.863Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {Prieto, Mario and Deus, Helena and Waard, Anita De and Schultes, Erik and García-Jiménez, Beatriz and Wilkinson, Mark D},
 doi = {10.7287/peerj.preprints.27829v1}
}

@inproceedings{
 title = {Modeling recovery of Crohn's disease, by simulating microbial community dynamics under perturbations},
 type = {inproceedings},
 year = {2019},
 websites = {https://doi.org/10.6084/m9.figshare.8864426,10.7490/f1000research.1117333,https://www.youtube.com/watch?v=xqUxwYZkn6w&list=PLmX8XnLr6zeHVOUbY0v6Jo22tUGpQSziV&index=5},
 city = {Basel},
 id = {d5484636-4b73-3797-9f2e-88082aae6a77},
 created = {2019-09-30T15:48:41.634Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2019-09-30T15:48:41.634Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {There are few large longitudinal microbiome studies, and fewer that include planned, annotated perturbations between sampling-points. Thus, there are few opportunities to employ data-driven computational analyses of perturbed microbial communities over time. Our novel computational system simulates the dynamics of microbial communities under perturbations, using genome-scale metabolic models (GEM). Perturbations include modifications to a) the nutrients available in the medium, allowing modelling of prebiotics; or b) the microorganisms present in the community, to model probiotics or pathogen infection. These simulations generate the quantity and types of information used as input to the MDPbiome system, which builds predictive models suggesting the perturbation(s) required to engineer microbial communities to a desired state. We demonstrate that this novel combination, called MDPbiomeGEM, is able to model the influence of prebiotic fiber and probiotic in the case of a Crohn's disease microbiome. The output's recommended perturbation to recover from dysbiosis is to consume inulin, which promotes butyrate production to reach homeostasis, consistent with prior biomed-ical knowledge. Our system could also contribute to design (perturbed) microbial community dynamics experiments, potentially saving resources both in natural microbiome scenarios by optimizing sequencing sampling, or to optimize in-vitro culture formulations for generating performant synthetic microbial communities.},
 bibtype = {inproceedings},
 author = {Carrasco Muriel, Jorge and García-Jiménez, Beatriz and Wlikinson, Mark D},
 doi = {10.13140/RG.2.2.33350.63049},
 booktitle = {Intelligent Systems for Molecular Biology and European Conference on Computational Biology}
}

There are few large longitudinal microbiome studies, and fewer that include planned, annotated perturbations between sampling-points. Thus, there are few opportunities to employ data-driven computational analyses of perturbed microbial communities over time. Our novel computational system simulates the dynamics of microbial communities under perturbations, using genome-scale metabolic models (GEM). Perturbations include modifications to a) the nutrients available in the medium, allowing modelling of prebiotics; or b) the microorganisms present in the community, to model probiotics or pathogen infection. These simulations generate the quantity and types of information used as input to the MDPbiome system, which builds predictive models suggesting the perturbation(s) required to engineer microbial communities to a desired state. We demonstrate that this novel combination, called MDPbiomeGEM, is able to model the influence of prebiotic fiber and probiotic in the case of a Crohn's disease microbiome. The output's recommended perturbation to recover from dysbiosis is to consume inulin, which promotes butyrate production to reach homeostasis, consistent with prior biomed-ical knowledge. Our system could also contribute to design (perturbed) microbial community dynamics experiments, potentially saving resources both in natural microbiome scenarios by optimizing sequencing sampling, or to optimize in-vitro culture formulations for generating performant synthetic microbial communities.

@inproceedings{
 title = {Condensed Microbiome Representation using Transfer and Deep Learning to Promote Microbial Composition Prediction},
 type = {inproceedings},
 year = {2019},
 websites = {https://doi.org/10.6084/m9.figshare.10336631},
 month = {11},
 day = {28},
 city = {Barcelona},
 id = {30daed69-f820-3c1b-9075-347c765d8340},
 created = {2019-12-18T11:52:18.840Z},
 accessed = {2019-12-18},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2019-12-18T11:52:18.919Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {Motivation:
Data produced by metagenomic studies has multiple layers of complexity.
Even 16S taxonomic analyses result in high-dimensional, extremely
complex data that thwarts knowledge discovery. In this study, we
describe a strategy to reduce the dimensionality of microbiome
datasets, such that they can be interrogated and explored more
easily.

Method:
This work brings together Deep Learning techniques, and microbiome data.
We selected a particular type of artificial neural network - an
autoencoder - to condense long vector values into a short vector
(i.e. an encoded representation) which is more amenable to various
kinds of analyses.  In this case, the long vector of values describes
a microbiome sample.  We further show that we are able to recover the
original vector from the encoded representation with high fidelity.

Results:
We transfer knowledge from a previously published dataset of around 5000
maize root microbiome samples into our autoencoder model, which
returns a code of 6 rational numbers representing the information
contained in the long vector of 717 taxa that describes the microbial
composition of those samples. We are subsequently able to predict 458
of those taxa, after decoding, with a Pearson correlation greater
than 0.5, with 0.77 being the average.  This compressed
representation opens-up many novel possibilities for microbiome data
analysis, particularly with respect to knowledge retrieval and
visualization. The autoencoder structure provides the ability to
recover the complete abundance vector from the codified samples,
making it possible to perform all analyses using the reduced coded
data, and to recover the long vector only when required. For example,
we apply our encoded microbiome to a novel scenario, showing that we
are able to predict the final microbial composition (717 taxa, after recovery of the original vector) of maize root microbiome samples using only a few available environmental variables such as plant age, temperature or precipitation. We achieve an average mean square error of 0.0018; this is a higher accuracy than predictions made without our encoded model.

Conclusions and Further work:
This condensed representation could be applied to any environment (gut,
ocean, urban soil, etc.) where there is a representative set of samples available. The contributions of our proposed microbiome autoencoder include: a) a novel dimensionality reduction approach to representing a long taxa vector as fewer than ten values; b) the ability to undertake challenging tasks in microbiome data analysis, such as to predict the microbial composition of hundreds of taxa based on a small number of features, rather than the more common (and
simpler) task of predicting a phenotypic feature of the
microbiome-associated host (e.g. age of the plant,  productivity or
disease) from hundreds of taxa; c) the encoded version of a
microbiome can be reused, via transfer learning, into novel but
related studies, allowing complex analyses to be undertaken using
fewer de novo sequencing samples; the knowledge encoded within
the microbiome autoencoder model can be applied to samples from a
similar environment, enabling inferences or predictions in studies
that would otherwise have insufficient power.},
 bibtype = {inproceedings},
 author = {Cabello, Sara and Garcia-Jimenez, Beatriz and Wilkinson, Mark D.},
 doi = {https://doi.org/10.6084/m9.figshare.10336631},
 booktitle = {Advances in Computational Biology Conference}
}

Motivation: Data produced by metagenomic studies has multiple layers of complexity. Even 16S taxonomic analyses result in high-dimensional, extremely complex data that thwarts knowledge discovery. In this study, we describe a strategy to reduce the dimensionality of microbiome datasets, such that they can be interrogated and explored more easily. Method: This work brings together Deep Learning techniques, and microbiome data. We selected a particular type of artificial neural network - an autoencoder - to condense long vector values into a short vector (i.e. an encoded representation) which is more amenable to various kinds of analyses. In this case, the long vector of values describes a microbiome sample. We further show that we are able to recover the original vector from the encoded representation with high fidelity. Results: We transfer knowledge from a previously published dataset of around 5000 maize root microbiome samples into our autoencoder model, which returns a code of 6 rational numbers representing the information contained in the long vector of 717 taxa that describes the microbial composition of those samples. We are subsequently able to predict 458 of those taxa, after decoding, with a Pearson correlation greater than 0.5, with 0.77 being the average. This compressed representation opens-up many novel possibilities for microbiome data analysis, particularly with respect to knowledge retrieval and visualization. The autoencoder structure provides the ability to recover the complete abundance vector from the codified samples, making it possible to perform all analyses using the reduced coded data, and to recover the long vector only when required. For example, we apply our encoded microbiome to a novel scenario, showing that we are able to predict the final microbial composition (717 taxa, after recovery of the original vector) of maize root microbiome samples using only a few available environmental variables such as plant age, temperature or precipitation. We achieve an average mean square error of 0.0018; this is a higher accuracy than predictions made without our encoded model. Conclusions and Further work: This condensed representation could be applied to any environment (gut, ocean, urban soil, etc.) where there is a representative set of samples available. The contributions of our proposed microbiome autoencoder include: a) a novel dimensionality reduction approach to representing a long taxa vector as fewer than ten values; b) the ability to undertake challenging tasks in microbiome data analysis, such as to predict the microbial composition of hundreds of taxa based on a small number of features, rather than the more common (and simpler) task of predicting a phenotypic feature of the microbiome-associated host (e.g. age of the plant, productivity or disease) from hundreds of taxa; c) the encoded version of a microbiome can be reused, via transfer learning, into novel but related studies, allowing complex analyses to be undertaken using fewer de novo sequencing samples; the knowledge encoded within the microbiome autoencoder model can be applied to samples from a similar environment, enabling inferences or predictions in studies that would otherwise have insufficient power.

@article{
 title = {MDPbiome: microbiome engineering through prescriptive perturbations},
 type = {article},
 year = {2018},
 pages = {i838-i847},
 volume = {34},
 websites = {https://academic.oup.com/bioinformatics/article/34/17/i838/5093255},
 month = {9},
 day = {1},
 id = {baa2f372-4158-367e-91e8-55b53d555041},
 created = {2019-09-30T15:06:24.528Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2019-09-30T15:53:05.855Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 abstract = {Motivation
Recent microbiome dynamics studies highlight the current inability to predict the effects of external perturbations on complex microbial populations. To do so would be particularly advantageous in fields such as medicine, bioremediation or industrial scenarios.

Results
MDPbiome statistically models longitudinal metagenomics samples undergoing perturbations as a Markov Decision Process (MDP). Given a starting microbial composition, our MDPbiome system suggests the sequence of external perturbation(s) that will engineer that microbiome to a goal state, for example, a healthier or more performant composition. It also estimates intermediate microbiome states along the path, thus making it possible to avoid particularly undesirable/unhealthy states. We demonstrate MDPbiome performance over three real and distinct datasets, proving its flexibility, and the reliability and universality of its output ‘optimal perturbation policy’. For example, an MDP created using a vaginal microbiome time series, with a goal of recovering from bacterial vaginosis, suggested avoidance of perturbations such as lubricants or sex toys; while another MDP provided a quantitative explanation for why salmonella vaccine accelerates gut microbiome maturation in chicks. This novel analytical approach has clear applications in medicine, where it could suggest low-impact clinical interventions that will lead to achievement or maintenance of a healthy microbial population, or alternately, the sequence of interventions necessary to avoid strongly negative microbiome states.},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and de la Rosa, Tomás and Wilkinson, Mark Denis},
 doi = {10.1093/bioinformatics/bty562},
 journal = {Bioinformatics},
 number = {17}
}

Motivation Recent microbiome dynamics studies highlight the current inability to predict the effects of external perturbations on complex microbial populations. To do so would be particularly advantageous in fields such as medicine, bioremediation or industrial scenarios. Results MDPbiome statistically models longitudinal metagenomics samples undergoing perturbations as a Markov Decision Process (MDP). Given a starting microbial composition, our MDPbiome system suggests the sequence of external perturbation(s) that will engineer that microbiome to a goal state, for example, a healthier or more performant composition. It also estimates intermediate microbiome states along the path, thus making it possible to avoid particularly undesirable/unhealthy states. We demonstrate MDPbiome performance over three real and distinct datasets, proving its flexibility, and the reliability and universality of its output ‘optimal perturbation policy’. For example, an MDP created using a vaginal microbiome time series, with a goal of recovering from bacterial vaginosis, suggested avoidance of perturbations such as lubricants or sex toys; while another MDP provided a quantitative explanation for why salmonella vaccine accelerates gut microbiome maturation in chicks. This novel analytical approach has clear applications in medicine, where it could suggest low-impact clinical interventions that will lead to achievement or maintenance of a healthy microbial population, or alternately, the sequence of interventions necessary to avoid strongly negative microbiome states.

@article{
 title = {FLYCOP: metabolic modeling-based analysis and engineering microbial communities},
 type = {article},
 year = {2018},
 pages = {i954-i963},
 volume = {34},
 websites = {https://academic.oup.com/bioinformatics/article/34/17/i954/5093244},
 month = {9},
 publisher = {Narnia},
 day = {1},
 id = {155e7a70-c635-30d4-965e-b8170fa13423},
 created = {2019-09-30T15:36:16.366Z},
 accessed = {2019-09-30},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2019-09-30T15:53:51.231Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {Motivation
Synthetic microbial communities begin to be considered as promising multicellular biocatalysts having a large potential to replace engineered single strains in biotechnology applications, in pharmaceutical, chemical and living architecture sectors. In contrast to single strain engineering, the effective and high-throughput analysis and engineering of microbial consortia face the lack of knowledge, tools and well-defined workflows. This manuscript contributes to fill this important gap with a framework, called FLYCOP (FLexible sYnthetic Consortium OPtimization), which contributes to microbial consortia modeling and engineering, while improving the knowledge about how these communities work. FLYCOP selects the best consortium configuration to optimize a given goal, among multiple and diverse configurations, in a flexible way, taking temporal changes in metabolite concentrations into account.

Results
In contrast to previous systems optimizing microbial consortia, FLYCOP has novel characteristics to face up to new problems, to represent additional features and to analyze events influencing the consortia behavior. In this manuscript, FLYCOP optimizes a Synechococcus elongatus-Pseudomonas putida consortium to produce the maximum amount of bio-plastic (PHA, polyhydroxyalkanoate), and highlights the influence of metabolites exchange dynamics in a four auxotrophic Escherichia coli consortium with parallel growth. FLYCOP can also provide an explanation about biological evolution driving evolutionary engineering endeavors by describing why and how heterogeneous populations emerge from monoclonal ones.},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and García, José Luis and Nogales, Juan},
 doi = {10.1093/bioinformatics/bty561},
 journal = {Bioinformatics},
 number = {17}
}

Motivation Synthetic microbial communities begin to be considered as promising multicellular biocatalysts having a large potential to replace engineered single strains in biotechnology applications, in pharmaceutical, chemical and living architecture sectors. In contrast to single strain engineering, the effective and high-throughput analysis and engineering of microbial consortia face the lack of knowledge, tools and well-defined workflows. This manuscript contributes to fill this important gap with a framework, called FLYCOP (FLexible sYnthetic Consortium OPtimization), which contributes to microbial consortia modeling and engineering, while improving the knowledge about how these communities work. FLYCOP selects the best consortium configuration to optimize a given goal, among multiple and diverse configurations, in a flexible way, taking temporal changes in metabolite concentrations into account. Results In contrast to previous systems optimizing microbial consortia, FLYCOP has novel characteristics to face up to new problems, to represent additional features and to analyze events influencing the consortia behavior. In this manuscript, FLYCOP optimizes a Synechococcus elongatus-Pseudomonas putida consortium to produce the maximum amount of bio-plastic (PHA, polyhydroxyalkanoate), and highlights the influence of metabolites exchange dynamics in a four auxotrophic Escherichia coli consortium with parallel growth. FLYCOP can also provide an explanation about biological evolution driving evolutionary engineering endeavors by describing why and how heterogeneous populations emerge from monoclonal ones.

@misc{
 title = {Memote: A community driven effort towards a standardized genome-scale metabolic model test suite},
 type = {misc},
 year = {2018},
 source = {bioRxiv},
 id = {50c018fd-9348-3bae-a815-5b7b0442fdc2},
 created = {2020-12-23T17:02:17.880Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2020-12-23T17:02:17.880Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 abstract = {The copyright holder for this preprint is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. Several studies have shown that neither the formal representation nor the functional requirements of genome-scale metabolic models (GEMs) are precisely defined. Without a consistent standard, comparability, reproducibility, and interoperability of models across groups and software tools cannot be guaranteed.Here, we present memote (https://github.com/opencobra/memote) an open-source software containing a community-maintained, standardized set of metabolic model tests. The tests cover a range of aspects from annotations to conceptual integrity and can be extended to include experimental datasets for automatic model validation. In addition to testing a model once, memote can be configured to do so automatically, i.e., while building a GEM. A comprehensive report displays the model’s performance parameters, which supports informed model development and facilitates error detection.Memote provides a measure for model quality that is consistent across reconstruction platforms and analysis software and simplifies collaboration within the community by establishing workflows for publicly hosted and version controlled models.},
 bibtype = {misc},
 author = {Lieven, C. and Beber, M.E. and Olivier, B.G. and Bergmann, F.T. and Ataman, M. and Babaei, P. and Bartell, J.A. and Blank, L.M. and Chauhan, S. and Correia, K. and Diener, C. and Dräger, A. and Ebert, B.E. and Edirisinghe, J.N. and Faria, J.P. and Feist, A. and Fengos, G. and Fleming, R.M.T. and García-Jiménez, B. and Hatzimanikatis, V. and van Helvoirt, W. and Henry, C.S. and Hermjakob, H. and Herrgård, M.J. and Kim, H.U. and King, Z. and Koehorst, J.J. and Klamt, S. and Klipp, E. and Lakshmanan, M. and Le Novère, N. and Lee, D.-Y. and Lee, S.Y. and Lee, S. and Lewis, N.E. and Ma, H. and Machado, D. and Mahadevan, R. and Maia, P. and Mardinoglu, A. and Medlock, G.L. and Monk, J.M. and Nielsen, J. and Nielsen, L.K. and Nogales, J. and Nookaew, I. and Resendis-Antonio, O. and Palsson, B.O. and Papin, J.A. and Patil, K.R. and Poolman, M. and Price, N.D. and Richelle, A. and Rocha, I. and Sanchez, B.J. and Schaap, P.J. and Malik Sheriff, R.S. and Shoaie, S. and Sonnenschein, N. and Teusink, B. and Vilaça, P. and Vik, J.O. and Wodke, J.A. and Xavier, J.C. and Yuan, Q. and Zakhartsev, M. and Zhang, C.},
 doi = {10.1101/350991}
}

The copyright holder for this preprint is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. Several studies have shown that neither the formal representation nor the functional requirements of genome-scale metabolic models (GEMs) are precisely defined. Without a consistent standard, comparability, reproducibility, and interoperability of models across groups and software tools cannot be guaranteed.Here, we present memote (https://github.com/opencobra/memote) an open-source software containing a community-maintained, standardized set of metabolic model tests. The tests cover a range of aspects from annotations to conceptual integrity and can be extended to include experimental datasets for automatic model validation. In addition to testing a model once, memote can be configured to do so automatically, i.e., while building a GEM. A comprehensive report displays the model’s performance parameters, which supports informed model development and facilitates error detection.Memote provides a measure for model quality that is consistent across reconstruction platforms and analysis software and simplifies collaboration within the community by establishing workflows for publicly hosted and version controlled models.

@techreport{
 title = {Temporal microbiome road-maps guided by perturbations},
 type = {techreport},
 year = {2016},
 websites = {http://biorxiv.org/lookup/doi/10.1101/049510},
 month = {4},
 day = {20},
 id = {b42a147b-23d9-3183-870d-c0f5398f959a},
 created = {2016-04-29T10:03:01.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {techreport},
 author = {García-Jiménez, Beatriz and Wilkinson, Mark},
 doi = {10.1101/049510}
}

@inproceedings{
 title = {MDPbiome: Predicting temporal microbiome dynamics influenced by external perturbations},
 type = {inproceedings},
 year = {2016},
 websites = {http://jbi2016.webs.upv.es/},
 city = {Valencia},
 id = {eb7c8253-8f59-31f1-a016-e12af177d0ba},
 created = {2016-04-29T10:10:34.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Wilkinson, Mark D.},
 booktitle = {XIII Symposium on Bioinformatics. JBI 2016. Poster}
}

@article{
 title = {Predicting Protein Relationships to Human Pathways through a Relational Learning Approach based on Simple Sequence Features},
 type = {article},
 year = {2014},
 pages = {753-765},
 volume = {11},
 websites = {http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6802366},
 id = {c6df837c-87f3-3d31-85bb-5f2a3bf6ca68},
 created = {2014-04-21T10:23:25.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and Pons, Tirso and Sanchis, Araceli and Valencia, Alfonso},
 doi = {10.1109/TCBB.2014.2318730},
 journal = {IEEE/ACM Transactions on Computational Biology and Bioinformatics},
 number = {4}
}

@inproceedings{
 title = {Automatic annotation of bioinformatics workflows with biomedical ontologies},
 type = {inproceedings},
 year = {2014},
 pages = {464-478},
 volume = {8803},
 publisher = {Springer-Verlag Heidelberg},
 series = {Lecture Notes in Computer Science},
 id = {5f3bcab8-0332-3a52-a6a2-caad45eb9da8},
 created = {2014-09-17T18:21:12.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 source_type = {CONF},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Wilkinson, Mark D},
 editor = {Margaria, T. and Steffen, B.},
 booktitle = {ISoLA2014: Proceedings of the 6th International Symposium On Leveraging Applications of Formal Methods, Verification and Validation}
}

@inproceedings{
 title = {Identifying Bioinformatics Sub-Workflows using Automated Biomedical Ontology Annotations},
 type = {inproceedings},
 year = {2014},
 websites = {http://www.bioinformaticsconference2014.org/},
 city = {Sevilla},
 id = {28820e27-55a5-31ca-b53f-5e755657700c},
 created = {2014-09-17T18:25:17.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Wilkinson, Mark D.},
 booktitle = {XII Symposium on Bioinformatics. JBI 2014. Poster}
}

@inbook{
 type = {inbook},
 year = {2014},
 pages = {119-156},
 websites = {http://goo.gl/UYG0o7},
 publisher = {CreateSpace},
 edition = {(In press)},
 chapter = {Minería de Datos},
 id = {49aaeb1f-b3d3-32d3-8b76-845a7fbc71b0},
 created = {2014-09-17T18:35:58.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 private_publication = {false},
 bibtype = {inbook},
 author = {García-Jiménez, Beatriz and Fernández, José María},
 editor = {Sebastián, Álvaro and Pascual-García, Alberto},
 title = {Bioinformática con Ñ}
}

@inproceedings{
 title = {The relevance of systems biology in relational learning to annotate human protein with Reactome pathways},
 type = {inproceedings},
 year = {2013},
 pages = {986},
 volume = {4},
 issue = {986},
 websites = {http://f1000.com/posters/browse/summary/1094191},
 month = {7},
 publisher = {F1000Posters},
 day = {19},
 city = {Berlin},
 series = {Protein Structure and Function Prediction and Analysis, poster L059},
 id = {92f549d0-214b-3182-9a2f-309329202746},
 created = {2014-03-27T15:10:52.000Z},
 accessed = {2014-03-27},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 country = {Germany},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Sanchis, Araceli},
 booktitle = {21st International Conference on Intelligent Systems for Molecular Biology and 12th European Conference on Computational Biology, ISMB/ECCB 2013}
}

@inproceedings{
 title = {Relational Learning-based Extension for Reactome Pathways with Sequence Features and Interactions},
 type = {inproceedings},
 year = {2012},
 websites = {http://jbi2012.org},
 month = {1},
 publisher = {Spanish Institute of Bioinformatics and Portuguese Bioinformatics Network},
 city = {Barcelona},
 id = {b19a7356-641d-3ffd-9563-219e995530eb},
 created = {2014-03-27T15:10:21.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 country = {Spain},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Pons, Tirso and Sanchis, Araceli and Valencia, Alfonso},
 booktitle = {11th Spanish Symposium on Bioinformatics. JBI 2012. Talk.}
}

@inproceedings{
 title = {MMRF for Proteome Annotation Applied to Human Protein Disease Prediction},
 type = {inproceedings},
 year = {2011},
 keywords = {First-Order Logic,Human Disease Annotation,Multi-Class Relational Decision Tree,Relational Data Mining,Structured Data},
 pages = {67-75},
 volume = {6489},
 websites = {http://dx.doi.org/10.1007/978-3-642-21295-6_11},
 month = {6},
 publisher = {Springer Berlin Heidelberg},
 city = {Florence},
 series = {Lecture Notes in Computer Science},
 id = {865b766e-122f-3163-900f-058926102d84},
 created = {2014-03-27T15:09:59.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 source_type = {CHAP},
 country = {Italy},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Ledezma, Agapito and Sanchis, Araceli},
 editor = {Frasconi, Paolo and Lisi, FrancescaA.},
 booktitle = {Proceedings of the 20th International Conference on Inductive Logic Programming, ILP 2010. Revised Papers.}
}

@article{
 title = {Inference of Functional Relations in Predicted Protein Networks with a Machine Learning Approach},
 type = {article},
 year = {2010},
 pages = {e9969},
 volume = {5},
 websites = {http://dx.doi.org/10.1371/journal.pone.0009969},
 month = {1},
 publisher = {Public Library of Science},
 id = {230220c5-fe5a-3a37-891b-3eb86bac9237},
 created = {2014-03-27T15:08:49.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 source_type = {JOUR},
 notes = {
        <m:bold>From Duplicate 1 ( </m:bold>
        <m:bold></m:bold>
        <m:bold>
          <m:italic>Inference of Functional Relations in Predicted Protein Networks with a Machine Learning Approach</m:italic>
        </m:bold>
        <m:bold></m:bold>
        <m:bold> - García-Jiménez, Beatriz; Juan, David; Ezkurdia, Iakes; Andrés-León, Eduardo; Valencia, Alfonso )<m:linebreak></m:linebreak>
        </m:bold>
        <m:linebreak></m:linebreak>M3: doi:10.1371/journal.pone.0009969<m:linebreak></m:linebreak>
        <m:linebreak></m:linebreak>
      },
 private_publication = {false},
 abstract = {Background Molecular biology is currently facing the challenging task of functionally characterizing the proteome. The large number of possible protein-protein interactions and complexes, the variety of environmental conditions and cellular states in which these interactions can be reorganized, and the multiple ways in which a protein can influence the function of others, requires the development of experimental and computational approaches to analyze and predict functional associations between proteins as part of their activity in the interactome. Methodology/Principal Findings We have studied the possibility of constructing a classifier in order to combine the output of the several protein interaction prediction methods. The AODE (Averaged One-Dependence Estimators) machine learning algorithm is a suitable choice in this case and it provides better results than the individual prediction methods, and it has better performances than other tested alternative methods in this experimental set up. To illustrate the potential use of this new AODE-based Predictor of Protein InterActions (APPIA), when analyzing high-throughput experimental data, we show how it helps to filter the results of published High-Throughput proteomic studies, ranking in a significant way functionally related pairs. Availability: All the predictions of the individual methods and of the combined APPIA predictor, together with the used datasets of functional associations are available at http://ecid.bioinfo.cnio.es/. Conclusions We propose a strategy that integrates the main current computational techniques used to predict functional associations into a unified classifier system, specifically focusing on the evaluation of poorly characterized protein pairs. We selected the AODE classifier as the appropriate tool to perform this task. AODE is particularly useful to extract valuable information from large unbalanced and heterogeneous data sets. The combination of the information provided by five prediction interaction prediction methods with some simple sequence features in APPIA is useful in establishing reliability values and helpful to prioritize functional interactions that can be further experimentally characterized.},
 bibtype = {article},
 author = {García-Jiménez, Beatriz and Juan, David and Ezkurdia, Iakes and Andrés-León, Eduardo and Valencia, Alfonso},
 journal = {PLoS ONE},
 number = {4}
}

Background Molecular biology is currently facing the challenging task of functionally characterizing the proteome. The large number of possible protein-protein interactions and complexes, the variety of environmental conditions and cellular states in which these interactions can be reorganized, and the multiple ways in which a protein can influence the function of others, requires the development of experimental and computational approaches to analyze and predict functional associations between proteins as part of their activity in the interactome. Methodology/Principal Findings We have studied the possibility of constructing a classifier in order to combine the output of the several protein interaction prediction methods. The AODE (Averaged One-Dependence Estimators) machine learning algorithm is a suitable choice in this case and it provides better results than the individual prediction methods, and it has better performances than other tested alternative methods in this experimental set up. To illustrate the potential use of this new AODE-based Predictor of Protein InterActions (APPIA), when analyzing high-throughput experimental data, we show how it helps to filter the results of published High-Throughput proteomic studies, ranking in a significant way functionally related pairs. Availability: All the predictions of the individual methods and of the combined APPIA predictor, together with the used datasets of functional associations are available at http://ecid.bioinfo.cnio.es/. Conclusions We propose a strategy that integrates the main current computational techniques used to predict functional associations into a unified classifier system, specifically focusing on the evaluation of poorly characterized protein pairs. We selected the AODE classifier as the appropriate tool to perform this task. AODE is particularly useful to extract valuable information from large unbalanced and heterogeneous data sets. The combination of the information provided by five prediction interaction prediction methods with some simple sequence features in APPIA is useful in establishing reliability values and helpful to prioritize functional interactions that can be further experimentally characterized.

@inproceedings{
 title = {S.cerevisiae Complex Function Prediction with Modular Multi-Relational Framework},
 type = {inproceedings},
 year = {2010},
 pages = {82-91},
 volume = {6098},
 websites = {http://link.springer.com/chapter/10.1007%2F978-3-642-13033-5_9},
 month = {6},
 publisher = {Springer},
 city = {Cordoba},
 series = {Lecture Notes in Artificial Intelligence},
 id = {0b923e30-1c77-32f7-9673-9c29a7af47c9},
 created = {2014-03-27T15:09:26.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 source_type = {CONF},
 country = {Spain},
 private_publication = {false},
 abstract = {Gene functions is an essential knowledge for understanding how metabolism works and designing treatments for solving malfunctions. The Modular Multi-Relational Framework (MMRF) is able to predict gene group functions. Since genes working together, it is focused on group functions rather than isolated gene functions. The approach of MMRF is flexible in several aspects, such as the kind of groups, the integration of different data sources, the organism and the knowledge representation. Besides, this framework takes advantages of the intrinsic relational structure of biological data, giving an easily biological interpretable and unique relational decision tree predicting N functions at once. This research work presents a group function prediction of S.cerevisiae (i.e.Yeast) genes grouped by protein complexes using MMRF. The results show that the predictions are restricted by the shortage of examples per class. Also, they assert that the knowledge representation is very determinant to exploit the available relational information richness, and therefore, to improve both the quantitative results and their biological interpretability.},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Ledezma Espino, Agapito and Sanchis de Miguel, Araceli},
 booktitle = {Trends in Applied Intelligent Systems: Proceedings of the 23rd International Conference on Industrial, Engineering & Other Applications of Applied Intelligent Systems (IEA-AIE 10)}
}

Gene functions is an essential knowledge for understanding how metabolism works and designing treatments for solving malfunctions. The Modular Multi-Relational Framework (MMRF) is able to predict gene group functions. Since genes working together, it is focused on group functions rather than isolated gene functions. The approach of MMRF is flexible in several aspects, such as the kind of groups, the integration of different data sources, the organism and the knowledge representation. Besides, this framework takes advantages of the intrinsic relational structure of biological data, giving an easily biological interpretable and unique relational decision tree predicting N functions at once. This research work presents a group function prediction of S.cerevisiae (i.e.Yeast) genes grouped by protein complexes using MMRF. The results show that the predictions are restricted by the shortage of examples per class. Also, they assert that the knowledge representation is very determinant to exploit the available relational information richness, and therefore, to improve both the quantitative results and their biological interpretability.

@article{
 title = {EcID. A Database for the Inference of Functional Interactions in E. coli},
 type = {article},
 year = {2009},
 pages = {629-635},
 volume = {37},
 websites = {http://nar.oxfordjournals.org/cgi/content/abstract/37/suppl_1/D629,http://nar.oxfordjournals.org/content/37/suppl_1/D629.abstract},
 id = {3e9ae025-1ec1-3c5d-99d4-fa21e9669734},
 created = {2014-03-27T15:07:34.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 source_type = {JOUR},
 notes = {
        <m:bold>From Duplicate 1 ( </m:bold>
        <m:bold>
          <m:italic>EcID. A database for the inference of functional interactions in E. coli</m:italic>
        </m:bold>
        <m:bold> - León, Eduardo Andrés; Ezkurdia, Iakes; García, Beatriz; Valencia, Alfonso; Juan, David )<m:linebreak></m:linebreak>
        </m:bold>
        <m:linebreak></m:linebreak>eprint: http://nar.oxfordjournals.org/cgi/reprint/37/suppl_1/D629.pdf<m:linebreak></m:linebreak>
        <m:linebreak></m:linebreak>
      },
 private_publication = {false},
 abstract = {The EcID database (Escherichia coli Interaction Database) provides a framework for the integration of information on functional interactions extracted from the following sources: EcoCyc (metabolic pathways, protein complexes and regulatory information), KEGG (metabolic pathways), MINT and IntAct (protein interactions). It also includes information on protein complexes from the two E. coli high-throughput pull-down experiments and potential interactions extracted from the literature using the web services associated to the iHOP text-mining system. Additionally, EcID incorporates results of various prediction methods, including two protein interaction prediction methods based on genomic information (Phylogenetic Profiles and Gene Neighbourhoods) and three methods based on the analysis of co-evolution (Mirror Tree, In Silico 2 Hybrid and Context Mirror). EcID associates to each prediction a specifically developed confidence score. The two main features that make EcID different from other systems are the combination of co-evolution-based predictions with the experimental data, and the introduction of E. coli-specific information, such as gene regulation information from EcoCyc. The possibilities offered by the combination of the EcID database information are illustrated with a prediction of potential functions for a group of poorly characterized genes related to yeaG. EcID is available online at http://ecid.bioinfo.cnio.es.},
 bibtype = {article},
 author = {León, Eduardo Andrés and Ezkurdia, Iakes and García-Jiménez, Beatriz and Valencia, Alfonso and Juan, David},
 journal = {Nucleic acids research},
 number = {suppl 1}
}

The EcID database (Escherichia coli Interaction Database) provides a framework for the integration of information on functional interactions extracted from the following sources: EcoCyc (metabolic pathways, protein complexes and regulatory information), KEGG (metabolic pathways), MINT and IntAct (protein interactions). It also includes information on protein complexes from the two E. coli high-throughput pull-down experiments and potential interactions extracted from the literature using the web services associated to the iHOP text-mining system. Additionally, EcID incorporates results of various prediction methods, including two protein interaction prediction methods based on genomic information (Phylogenetic Profiles and Gene Neighbourhoods) and three methods based on the analysis of co-evolution (Mirror Tree, In Silico 2 Hybrid and Context Mirror). EcID associates to each prediction a specifically developed confidence score. The two main features that make EcID different from other systems are the combination of co-evolution-based predictions with the experimental data, and the introduction of E. coli-specific information, such as gene regulation information from EcoCyc. The possibilities offered by the combination of the EcID database information are illustrated with a prediction of potential functions for a group of poorly characterized genes related to yeaG. EcID is available online at http://ecid.bioinfo.cnio.es.

@inproceedings{
 title = {Modular Multi-Relational Framework for Gene Group Function Prediction.},
 type = {inproceedings},
 year = {2009},
 websites = {http://dtai.cs.kuleuven.be/ilp-mlg-srl/USBStick/papers/ILP09-52.pdf},
 month = {7},
 publisher = {Springer},
 city = {Leuven},
 id = {ee48cb30-5ca9-3366-ad1f-e59d27235b07},
 created = {2014-03-27T15:08:18.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 source_type = {CONF},
 country = {Belgium},
 notes = {http://www.cs.kuleuven.be/~dtai/ilp-mlg-srl/papers/ILP09-52.pdf},
 private_publication = {false},
 abstract = {Determining the functions of genes is essential for understanding how the metabolisms work, and for trying to solve their malfunctions. Genes usually work in groups rather than isolated, so functions should be assigned to gene groups and not to individual genes. Moreover, the genetic knowledge has many relations and is very frequently changeable. Thus, a propositional ad-hoc approach is not appropriate to deal with the gene group function prediction domain. We propose the Modular Multi-Relational Framework (MMRF), which faces the problem from a relational and flexible point of view. The MMRF consists of several modules covering all involved domain tasks (grouping, representing and learning using computational prediction techniques). A specific application is described, including a relational representation language, where each module of MMRF is individually instantiated and refined for obtaining a prediction under specific given conditions.},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Ledezma, Agapito and Sanchis, Araceli},
 booktitle = {ILP '09: 19th International Conference on Inductive Logic Programming}
}

Determining the functions of genes is essential for understanding how the metabolisms work, and for trying to solve their malfunctions. Genes usually work in groups rather than isolated, so functions should be assigned to gene groups and not to individual genes. Moreover, the genetic knowledge has many relations and is very frequently changeable. Thus, a propositional ad-hoc approach is not appropriate to deal with the gene group function prediction domain. We propose the Modular Multi-Relational Framework (MMRF), which faces the problem from a relational and flexible point of view. The MMRF consists of several modules covering all involved domain tasks (grouping, representing and learning using computational prediction techniques). A specific application is described, including a relational representation language, where each module of MMRF is individually instantiated and refined for obtaining a prediction under specific given conditions.

@inproceedings{
 title = {Protein-protein functional association prediction using genetic programming},
 type = {inproceedings},
 year = {2008},
 pages = {347-348},
 websites = {http://dl.acm.org/citation.cfm?id=1389095.1389156},
 month = {7},
 publisher = {ACM Press},
 day = {13},
 city = {Atlanta},
 id = {0a3c763c-a471-3a72-a0b5-ed92ed492423},
 created = {2014-03-27T15:06:31.000Z},
 accessed = {2014-03-27},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 country = {USA},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Aler, Ricardo and Ledezma, Agapito and Sanchis, Araceli},
 booktitle = {Proceedings of the 10th annual conference on Genetic and Evolutionary Computation - GECCO '08}
}

@inproceedings{
 title = {Genetic Programming for Predicting Protein Networks},
 type = {inproceedings},
 year = {2008},
 keywords = {bioinformatics,classification,data integra-,evolutionary computation,genetic programming,machine learning,protein interaction prediction,tion},
 pages = {432-441},
 volume = {5290},
 websites = {http://link.springer.com/chapter/10.1007%2F978-3-540-88309-8_44},
 month = {10},
 publisher = {Springer},
 city = {Lisbon},
 series = {Lecture Notes in Artificial Intelligence. Advances in Artificial Intelligence},
 id = {f120d68e-0915-3940-a4ac-a6bde266f013},
 created = {2014-03-27T15:06:44.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 country = {Portugal},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {García-Jiménez, Beatriz and Aler, Ricardo and Ledezma, Agapito and Sanchis, Araceli},
 editor = {Geffner, Hector},
 booktitle = {Proceedings of the 11th Ibero-American Conference on Artificial Intelligence (IBERAMIA'08)}
}

@inproceedings{
 title = {Studying the Capacity of Grammatical Encoding to Generate FNN Architectures},
 type = {inproceedings},
 year = {2003},
 pages = {478-485},
 websites = {http://link.springer.com/chapter/10.1007/3-540-44868-3_61#},
 month = {6},
 publisher = {Springer},
 city = {Maó, Menorca},
 series = {Lecture Notes in Computer Science},
 id = {f7c42f90-599c-36f8-9dcf-454eda4104f3},
 created = {2014-03-27T15:06:14.000Z},
 file_attached = {false},
 profile_id = {fa910c8b-8889-3a42-afc9-302da7e3933a},
 group_id = {ff1f9038-dd83-321a-9605-910d757253bb},
 last_modified = {2017-03-14T12:53:44.413Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 country = {Spain},
 private_publication = {false},
 bibtype = {inproceedings},
 author = {Gutiérrez, Germán and García-Jiménez, Beatriz and Molina, José M. and Sanchis, Araceli},
 editor = {Mira, José and Álvarez, José R.},
 booktitle = {Computational Methods in Neural Modeling Proceedings of the 7th Work-Conference on Artificial and Natural Neuronal Networks (IWANN 2003)}
}