| layout | title | permalink |
|---|---|---|
page |
Overview |
/overview/ |
This page outlines the resources generated for the Human Brain Pharmacome project lead by Charles Tapley Hoyt in 2018 and 2019.
One of the first goals of the Human Brain Pharmacome Project is to bring order to the unstructured knowledge locked in the biomedical literature, patents, and electronic health records.
We heavily rely on Biological Expression Language (BEL) as a format for storing qualitative causal and correlative relations between biological entities across multiple modes and scales, with full provenance information including namespace references, relation provenance (citation and evidence), and biological context-specific relation metadata (anatomy, cell, disease etc.)
Slater, T. (2014). Recent advances in modeling languages for pathway maps and computable biological networks. Drug Discovery Today, 19(2), 193–198.
- BEL Enhancement Proposals: https://github.com/belbio/bep
We built
PyBEL for parsing, validating, compiling, and converting networks encoded in BEL. PyBEL comprises five main components:
(i) network data container, (ii) parser and validator, (iii) network database manager, (iv) data converter and (v)
network visualizer.
Hoyt, C. T., Konotopez, A., Ebeling, C. (2018). PyBEL: a computational framework for Biological Expression Language. Bioinformatics, 34(4), 703–704.
Hoyt, C. T., Domingo-Fernández, D., & Hofmann-Apitius, M. (2018). BEL Commons: an environment for exploration and analysis of networks encoded in Biological Expression Language. Database, 2018(3), 1–11.
- Web Application: https://bel-commons.scai.fraunhofer.de
- Code
We developed a workflow using git, GitHub, PyBEL, and a novel PyBEL extension in order to identify and address syntactical issues in BEL documents in a Continuous Integration setting.
- Source Code
- Zenodo Reference: Hoyt, C. T. (2018, December 22). pybel/pybel-git v0.0.2 (Version v0.0.2). Zenodo. http://doi.org/10.5281/zenodo.2508180
While there are several useful public terminologies useful for curation of biomedical relations, there is often the need to develop new controlled vocabularies, thesauri, taxonomies, and ontologies. We have maintained ours with the ability to export it as OBO, OWL, and as a BEL namespace.
We have prioritized manual curation of the highest granularity for new content related to several neurodegenerative disease, with special focus on the human Tau protein, nicotoinic receptor biology, and proteostasis. They can be downloaded and handled with PyBEL (or other BEL tools) using the link below.
Domingo-Fernández, D., et al. (2017). Multimodal mechanistic signatures for neurodegenerative diseases (NeuroMMSig): A web server for mechanism enrichment. Bioinformatics, 33(22), 3679–3681.
- Web Application: https://neurommsig.fraunhofer.de
Hoyt, C. T., et al (2019). Re-curation and Rational Enrichment of Knowledge Graphs in Biological Expression Language. Database, 2019.
We used the tooling developed for the rational enrichment workflow to support topic-specific, semi-automated curation. So far, it has been applied to the MAPT and GSK3B proteins.
We've generated a dynamic web application for exploring the content curated during this project that can be found at https://github.com/pharmacome/taubase. It will be deployed publicly soon.
Hoyt, C. T., et al. (2019). Integration of Structured Biological Data Sources using Biological Expression Language. bioRxiv, 631812.
Domingo-Fernandez, D., et al. (2018). ComPath: an ecosystem for exploring, analyzing, and curating mappings across pathway databases. Npj Systems Biology and Applications, 4:43.
- Web Application: https://compath.scai.fraunhofer.de
- Code
Domingo-Fernandez, D., et al. (2018). PathMe: Merging and exploring mechanistic pathway knowledge. BMC Bioinformatics, 20:243.
- Web Application: https://pathme.scai.fraunhofer.de
- Code
We generated super-pathways using ComPath and PathMe in order to benchmark several pathway-based functional enrichment and prediction methods.
Mubeen, S., et al. (2019). The Impact of Pathway Database Choice on Statistical Enrichment Analysis and Predictive Modeling. Frontiers in Genetics, 10:1203.
We proposed an explanation for the cross-indication effects of Carbamazepine towards epilepsy and Alzheimer's disease by using the NeuroMMSig mechanism enrichment server to identify overlapping mechanisms with its experimentally measured and computationally predicted targets.
Hoyt, C. T. and Domingo-Fernández, D., et al. (2018). A systematic approach for identifying shared mechanisms in epilepsy and its comorbidities. Database, 2018(1).
We used GAT2VEC to generate random walk-based node embeddings for protein-protein interaction (PPI) networks annotated with disease-specific differential gene expression profiles from GEO, ArrayExpress, and other sources. Following the prediction and evaluation pipeline from Emig, et al. (2013), we built simple generalized linear models using annotations from OpenTargets as a training set and generated rankings for all targets.
Muslu, Ö., Hoyt, C. T., Hofmann-Apitius, M., & Fröhlich, H. (2019). GuiltyTargets: Prioritization of Novel Therapeutic Targets with Deep Network Representation Learning. bioRxiv, 1–14.
We generated two network representation learning (NRL) libraries: one for random walk-based NRL and another for semantic translations models and neural network models.
Ali, M., Hoyt, C. T., Domingo-Fernández, D., Lehmann, J., & Jabeen, H. (2019). BioKEEN: A library for learning and evaluating biological knowledge graph embeddings. Bioinformatics, btz117.
We generated node and edge embeddings for a graph of drugs, their chemical similarities, their side effects, their indications, their targets in order to predict side effects for new chemicals as well as to propose mechanisms of action for some side effects.
- Code
- master's thesis in preparation
We replaced the engineered features used by Himmelstein et al. (2017) with features learned by node2vec on the RepHetNet.
- Code
- master's thesis in preparation
We're funded by the Fraunhofer Society's MAVO program as a joint undertaking between Fraunhofer SCAI, Fraunhofer IAIS, and Fraunhofer IME. Also, see our blog.