Biology subjects

Bachman, J. A.

Publications and source records attributed to Bachman, J. A..

9 recordsLinked to original sources

bioRxiv2022

A versatile and interoperable computational framework for the analysis and modeling of COVID-19 disease mechanisms

The COVID-19 Disease Map project is a large-scale community effort uniting 277 scientists from 130 Institutions around the globe. We use high-quality, mechanistic content describing SARS-CoV-2-host interactions and develop interoperable bioinformatic pipelines for novel target identification and drug repurposing. Community-driven and highly interdisciplinary, the project is collaborative and supports community standards, open access, and the FAIR data principles. The coordination of community work allowed for an impressive step forward in building interfaces between Systems Biology tools and platforms. Our framework links key molecules highlighted from broad omics data analysis and computational modeling to dysregulated pathways in a cell-, tissue- or patient-specific manner. We also employ text mining and AI-assisted analysis to identify potential drugs and drug targets and use topological analysis to reveal interesting structural features of the map. The proposed framework is versatile and expandable, offering a significant upgrade in the arsenal used to understand virus-host interactions and other complex pathologies.

systems biology↗

bioRxiv2022

Automated assembly of molecular mechanisms at scale from text mining and curated databases

The analysis of omic data depends heavily on machine-readable information about protein interactions, modifications, and activities. Key resources include protein interaction networks, databases of post-translational modifications, and curated models of gene and protein function. Software systems that read primary literature can potentially extend and update such resources while reducing the burden on human curators, but machine-reading software systems have a high error rate. Here we describe an approach to precisely assemble molecular mechanisms at scale using natural language processing systems and the Integrated Network and Dynamical Reasoning Assembler (INDRA). INDRA identifies overlaps and redundancies in information extracted from published papers and pathway databases and uses probability models to reduce machine reading errors. INDRA enables the automated creation of high-quality, non-redundant corpora for use in data analysis and causal modeling. We demonstrate the use of INDRA in extending protein-protein interaction databases and explaining co-dependencies in the Cancer Dependency Map.

systems biology↗

bioRxiv2021

STonKGs: A Sophisticated Transformer Trained on Biomedical Text and Knowledge Graphs

The majority of biomedical knowledge is stored in structured databases or as unstructured text in scientific publications. This vast amount of information has led to numerous machine learning-based biological applications using either text through natural language processing (NLP) or structured data through knowledge graph embedding models (KGEMs). However, representations based on a single modality are inherently limited. To generate better representations of biological knowledge, we propose STonKGs, a Sophisticated Transformer trained on biomedical text and Knowledge Graphs. This multimodal Transformer uses combined input sequences of structured information from KGs and unstructured text data from biomedical literature to learn joint representations. First, we pre-trained STonKGs on a knowledge base assembled by the Integrated Network and Dynamical Reasoning Assembler (INDRA) consisting of millions of text-triple pairs extracted from biomedical literature by multiple NLP systems. Then, we benchmarked STonKGs against two baseline models trained on either one of the modalities (i.e., text or KG) across eight different classification tasks, each corresponding to a different biological application. Our results demonstrate that STonKGs outperforms both baselines, especially on the more challenging tasks with respect to the number of classes, improving upon the F1-score of the best baseline by up to 0.083. Additionally, our pre-trained model as well as the model architecture can be adapted to various other transfer learning applications. Finally, the source code and pre-trained STonKGs models are available at https://github.com/stonkgs/stonkgs and https://huggingface.co/stonkgs/stonkgs-150k.

bioinformatics↗

bioRxiv2021

Capturing scientific knowledge in computable form

Technological advances in computing provide major opportunities to accelerate scientific discovery. The wide availability of structured knowledge would allow us to take full advantage of these by enabling efficient human-computer interaction. Traditionally, biological knowledge is captured in publications and knowledge bases, however, the information in articles is not directly accessible to computers, and knowledge bases are constrained by finite resources available for manual curation. To accelerate knowledge capture and communication and to keep pace with the rapid growth of scientific reports, we developed the Biofactoid (biofactoid.org) software suite, which crowdsources structured knowledge in articles from authors. Biofactoid is a web-based system that lets scientists draw a network of interactions between genes, their products, and chemical compounds and employs smart-automation to translate user input into a structured language using the expressive power of a formal ontology. The resulting data is shared via public information resources, enabling author-curated knowledge to be appreciated in the context of all existing computable knowledge. Authors of recently published papers across a range of journals have already contributed their pathway information, much of which is novel and extends existing pathway databases into new biological areas. We envision the adoption of Biofactoid for crowdsourced curation by scientists and publishers as part of an ecosystem of tools that accelerate scientific communication and discovery. AvailabilityBiofactoid server at https://biofactoid.org

bioinformatics↗

bioRxiv2020

Exploring the understudied human kinome for research and therapeutic opportunities

The functions of protein kinases have been widely studied and over 60 kinase inhibitors are FDA-approved drugs. Membership in the human kinome is nonetheless subject to multiple overlapping and inconsistent definitions and is unevenly studied, complicating functional genomics and chemical genetics. We describe objective criteria for refining the definition of the human kinome to comprise an extended set of 710 kinase domains and a more narrowly curated set of 557 protein kinase like (PKL) domains. An online tool (www.kinome.org) makes it possible to sort these sets on multiple structural and functional criteria. Focusing on the least studied one-third of the kinome we find that many proteins are differentially expressed, essential in multiple cell lines, and mutated in the Cancer Genome Atlas. We show that some understudied kinases are high affinity off-targets of clinical-grade compounds and approved drugs and we describe an optimized small molecule library making use of this information for selective kinome perturbation. We conclude that the understudied kinome contains physiologically important proteins, including possible targets for future drug discovery campaigns.

systems biology↗

bioRxiv2019

Assembling a corpus of phosphoproteomic annotations using ProtMapper to normalize site information from databases and text mining

Protein phosphorylation regulates numerous cellular processes and is highly studied in biology.However, the analysis of phosphoproteomic datasets remains challenging due to limited information on upstream regulators of phosphosites, which is fragmented across multiple curated databases and unstructured literature. When aggregating information on phosphosites from six databases and three text mining systems, we found that a substantial proportion of phosphosites were mentioned at residue positions not matching the reference sequence. These errors were often attributable to the use of residue numbers from non-canonical protein isoforms, mouse or rat proteins, or post-translationally processed proteins. Non-canonical site numbering is also prevalent in mass spectrometry datasets from large-scale efforts such as the Clinical Proteomic Tumor Analysis Consortium (CPTAC). To address these issues, we developed ProtMapper, an open-source Python tool that automatically normalizes site positions to human protein reference sequences. We used ProtMapper coupled with the INDRA knowledge assembly system to create a corpus of 37,028 regulatory annotations for 16,332 sites - to our knowledge, the most comprehensive corpus of literature-derived information about phosphosite regulation currently available. This work highlights how automated phosphosite normalization coupled to text mining and knowledge assembly allows researchers to leverage phosphosite information that exists within the scientific literature.

bioinformatics↗

bioRxiv2019

GeneWalk identifies relevant gene functions for a biological context using network representation learning

The primary bottleneck in high-throughput genomics experiments is identifying the most important genes and their relevant functions from a list of gene hits. Existing methods such as Gene Ontology (GO) enrichment analysis provide insight at the gene set level. For individual genes, GO annotations are static and biological context can only be added by manual literature searches. Here, we introduce GeneWalk (github.com/churchmanlab/genewalk), a method that identifies individual genes and their relevant functions under a particular experimental condition. After automatic assembly of an experiment-specific gene regulatory network, GeneWalk quantifies the similarity between vector representations of each gene and its GO annotations through representation learning, yielding annotation significance scores that reflect their functional relevance for the experimental context. We demonstrate the use of GeneWalk analysis of RNA-seq and nascent transcriptome (NET-seq) data from human cells and mouse brains, validating the methodology. By performing gene- and condition-specific functional analysis that converts a list of genes into data-driven hypotheses, GeneWalk accelerates the interpretation of high-throughput genetics experiments.

bioinformatics↗

bioRxiv2017

Bioentities: a resource for entity recognition and relationship resolution in biomedical text mining

BackgroundFor automated reading of scientific publications to extract useful information about molecular mechanisms it is critical that genes, proteins and other entities be correctly associated with uniform identifiers, a process known as named entity linking or \"grounding.\" Correct grounding is essential for resolving relationships among mined information, curated interaction databases, and biological datasets. The accuracy of this process is largely dependent on the availability of machine-readable resources associating synonyms and abbreviations commonly found in biomedical literature with uniform identifiers.\n\nResultsIn a task involving automated reading of [~]215,000 articles using the REACH event extraction software we found that grounding was disproportionately inaccurate for multi-protein families (e.g., \"AKT\") and complexes with multiple subunits (e.g.\"NF-{kappa}B\"). To address this problem we constructed FamPlex, a manually curated resource defining protein families and complexes as they are commonly encountered in biomedical text. In FamPlex the gene-level constituents of families and complexes are defined in a flexible format allowing for multi-level, hierarchical membership. To create FamPlex, text strings corresponding to entities were identified empirically from literature and linked manually to uniform identifiers; these identifiers were also mapped to equivalent entries in multiple related databases. FamPlex also includes curated prefix and suffix patterns that improve named entity recognition and event extraction. Evaluation of REACH extractions on a test corpus of [~]54,000 articles showed that FamPlex significantly increased grounding accuracy for families and complexes (from 15% to 71%). The hierarchical organization of entities in FamPlex also made it possible to integrate otherwise unconnected mechanistic information across families, subfamilies, and individual proteins. Applications of FamPlex to the TRIPS/DRUM reading system and the Biocreative VI Bioentity Normalization Task dataset demonstrated the utility of FamPlex in other settings.\n\nConclusionFamPlex is an effective resource for improving named entity recognition, grounding, and relationship resolution in automated reading of biomedical text. The content in FamPlex is available in both tabular and Open Biomedical Ontology formats at https://github.com/sorgerlab/famplex under the Creative Commons CC0 license and has been integrated into the TRIPS/DRUM and REACH reading systems.

bioinformatics↗

bioRxiv2017

From word models to executable models of signaling networks using automated assembly

Word models (natural language descriptions of molecular mechanisms) are a common currency in spoken and written communication in biomedicine but are of limited use in predicting the behavior of complex biological networks. We present an approach to building computational models directly from natural language using automated assembly. Molecular mechanisms described in simple English are read by natural language processing algorithms, converted into an intermediate representation and assembled into executable or network models. We have implemented this approach in the Integrated Network and Dynamical Reasoning Assembler (INDRA), which draws on existing natural language processing systems as well as pathway information in Pathway Commons and other online resources. We demonstrate the use of INDRA and natural language to model three biological processes of increasing scope: (i) p53 dynamics in response to DNA damage; (ii) adaptive drug resistance in BRAF-V600E mutant melanomas; and (iii) the RAS signaling pathway. The use of natural language for modeling makes routine tasks more efficient for modeling practitioners and increases the accessibility and transparency of models for the broader biology community.\n\nStandfirst textINDRA uses natural language processing systems to read descriptions of molecular mechanisms and assembles them into executable models.\n\nHighlightsO_LIINDRA decouples the curation of knowledge as word models from model implementation\nC_LIO_LIINDRA is connected to multiple natural language processing systems and can draw on information from curated databases\nC_LIO_LIINDRA can assemble dynamical models in rule-based and reaction network formalisms, as well as Boolean networks and visualization formats\nC_LIO_LIWe used INDRA to build models of p53 dynamics, resistance to targeted inhibitors of BRAF in melanoma, and the Ras signaling pathway from natural language\nC_LI

systems biology↗