https://doi.org/10.1021/acs.jcim.5c02297

2025-11-14

2025-11-14

https://doi.org/10.26434/chemrxiv-2025-xr5h2

2025-09-30

2025-09-30

This poster presents results from a data integration framework that combines clinical trial metadata to support drug target hypothesis generation through evidence aggregation. We apply a cross-mapping approach across biomedical, chemical, and biological data sources—including the ClinicalTrials.gov AACT database—to infer disease–target associations. The methodology involves mapping drugs to their protein-coding gene targets, linking them to diseases, and scoring these associations using statistical metrics that reflect evidence. This structured approach enables interpretable and reproducible insights in support of the Illuminating the Druggable Genome (IDG) initiative. Presented at the DE Shaw Research Fellowship Meeting, New York, NY, April 10, 2025.

2025-05-07

2025-05-07

This poster presents results from a data integration framework that combines clinical trial metadata to support drug target hypothesis generation through evidence aggregation. We apply a cross-mapping approach across biomedical, chemical, and biological data sources—including the ClinicalTrials.gov AACT database—to infer disease–target associations. The methodology involves mapping drugs to their protein-coding gene targets, linking them to diseases, and scoring these associations using statistical metrics that reflect evidence. This structured approach enables interpretable and reproducible insights in support of the Illuminating the Druggable Genome (IDG) initiative. Presented at the DE Shaw Research Fellowship Meeting, New York, NY, April 10, 2025.

2025-05-07

2025-05-07

Objective: To develop a machine learning (ML) pipeline for identifying potential gene-disease associations, even when explicit connections are absent in knowledge graphs. Background: Biomedical knowledge graphs (KGs), such as the Data Distillery Knowledge Graph (DDKG), capture known relationships among entities (e.g., genes, diseases, proteins), providing valuable insights for research. However, these relationships are typically derived from prior studies, leaving potential unknown associations unexplored. Traditional methods, like linkage analysis and genome-wide association studies (GWAS), can be time-consuming and resource-intensive. Recently, network-based methods and KGs, enhanced by advances in ML frameworks, have emerged as powerful tools for inferring these unexplored associations. Methods: We propose KG2ML (Knowledge Graph to Machine Learning), a pipeline utilizing the PULSNAR algorithm (Positive Unlabeled Learning Selected Not At Random) to infer disease-associated genes not explicitly recorded in the DDKG. CondensedKG, a subset of DDKG with 8 node types and 1,042 relationship types, was created to improve data interpretability. KG2ML transforms CondensedKG into a feature matrix using ProteinGraphML’s path-based method, and then PULSNAR generates probabilistic scores for gene-disease associations. KG2ML was applied to 12 diseases, including Bipolar Disorder, Coronary Artery Disease, and Parkinson's Disease. The top 15 high-probability genes per disease were validated through expert review and reference databases (e.g., TINX). XGBoost models were evaluated with 5-fold cross-validation to measure the impact of incorporating PULSNAR-imputed genes as positives. Results: For some of those 12 diseases, 14 of the 15 top-ranked genes lacked prior explicit associations but were supported by literature and TINX evidence. Incorporating PULSNAR-imputed genes as positives enhanced XGBoost classification, highlighting PU learning's potential in identifying hidden gene-disease relationships. Conclusions: KG2ML demonstrates that PU learning can effectively discover disease-gene associations missing from current KGs. By combining KG data with ML-based inference, it offers a scalable, interpretable strategy to advance biomedical research and address KG limitations.

2025-04-18

2025-04-18

Objective: To develop a machine learning (ML) pipeline for identifying potential gene-disease associations, even when explicit connections are absent in knowledge graphs. Background: Biomedical knowledge graphs (KGs), such as the Data Distillery Knowledge Graph (DDKG), capture known relationships among entities (e.g., genes, diseases, proteins), providing valuable insights for research. However, these relationships are typically derived from prior studies, leaving potential unknown associations unexplored. Traditional methods, like linkage analysis and genome-wide association studies (GWAS), can be time-consuming and resource-intensive. Recently, network-based methods and KGs, enhanced by advances in ML frameworks, have emerged as powerful tools for inferring these unexplored associations. Methods: We propose KG2ML (Knowledge Graph to Machine Learning), a pipeline utilizing the PULSNAR algorithm (Positive Unlabeled Learning Selected Not At Random) to infer disease-associated genes not explicitly recorded in the DDKG. CondensedKG, a subset of DDKG with 8 node types and 1,042 relationship types, was created to improve data interpretability. KG2ML transforms CondensedKG into a feature matrix using ProteinGraphML’s path-based method, and then PULSNAR generates probabilistic scores for gene-disease associations. KG2ML was applied to 12 diseases, including Bipolar Disorder, Coronary Artery Disease, and Parkinson's Disease. The top 15 high-probability genes per disease were validated through expert review and reference databases (e.g., TINX). XGBoost models were evaluated with 5-fold cross-validation to measure the impact of incorporating PULSNAR-imputed genes as positives. Results: For some of those 12 diseases, 14 of the 15 top-ranked genes lacked prior explicit associations but were supported by literature and TINX evidence. Incorporating PULSNAR-imputed genes as positives enhanced XGBoost classification, highlighting PU learning's potential in identifying hidden gene-disease relationships. Conclusions: KG2ML demonstrates that PU learning can effectively discover disease-gene associations missing from current KGs. By combining KG data with ML-based inference, it offers a scalable, interpretable strategy to advance biomedical research and address KG limitations.

2025-04-18

2025-04-18

Poster presented at the NIH Common Fund Data Ecosystem (CFDE) All-Hands Meeting in Bethesda, MD, March 19-20, 2024.

2025-05-01

2025-05-01

Poster presented at the NIH Common Fund Data Ecosystem (CFDE) All-Hands Meeting in Bethesda, MD, March 19-20, 2024.

2025-05-01

2025-05-01

https://doi.org/10.1101/2025.03.17.25323906

2025-03-20

2025-10-23

https://doi.org/10.1109/JBHI.2024.3515805

2025-02-10

2025-02-10

https://doi.org/10.1101/2025.01.23.25321034

2025-01-26

2025-01-26

https://doi.org/10.7717/peerj.17470

2024-06-25

2024-06-25

Poster presented at the NIH Common Fund Data Ecosystem (CFDE) All-Hands Meeting in Bethesda, MD, March 19-20, 2024.

2025-04-18

2025-04-18

Poster presented at the NIH Common Fund Data Ecosystem (CFDE) All-Hands Meeting in Bethesda, MD, March 19-20, 2024.

2025-04-18

2025-04-18

https://doi.org/10.1093/bioadv/vbae036

2024-04-05

2024-04-05

Invited presentation of the UNM Biomedical Seminar Series (BioMISS) organized by the UNM Health Sciences Library and Informatics Center (HSLIC). This seminar describes several biomedical data science research projects from diverse domains, involving teams comprised of contributors from UNM and elsewhere, which share a common theme of evidence evaluation for pharmaceutical discovery. What is the strongest biomedical evidence about a disease for discovery of novel pharmaceutical therapies? This is a fundamental challenge for biomedical scientists, but also translates to a parallel question for data science: Can we systematically assemble and query biomedical knowledge graphs in a computational discovery platform guided by rational, algorithmic measures of relevance and confidence, facilitating scientific discovery? And, how have continuing waves of scientific and technological progress, in an era of bigger and bigger data, informed and empowered these inquiries?

2023-10-07

2023-10-07

https://doi.org/10.1002/ntls.20220067

2023-05-01

2023-07-05

https://doi.org/10.1093/nar/gkac1033

2022-11-29

2023-01-08

https://doi.org/10.1093/bioinformatics/btab427

2021-07-23

2022-06-01

Talk presented at the ISMB-ECCB Bioinformatics Open Source Conference (BOSC2021), July 25-30, 2021.

2022-01-05

2022-06-02

Talk presented at the ISMB-ECCB Bioinformatics Open Source Conference (BOSC2021), July 25-30, 2021.

2022-01-05

2022-06-02

https://doi.org/10.1101/2020.12.30.424881

2021-01-12

2024-08-05

https://doi.org/10.1101/2020.11.11.378596

2021-01-12

2024-08-05

GTEx-based expression profiles, sex as a biological variable (SABV) analysis. This plot compares metrics for pairwise distance/similarity between profiles, represented as real valued vectors.

2018-10-26

2022-05-27

GTEx-based expression profiles, sex as a biological variable (SABV) analysis. This plot compares metrics for pairwise distance/similarity between profiles, represented as real valued vectors.

2019-02-25

2022-05-27

Shows how the Badapple formula for compound scaffold promiscuity combines weight of evidence with bioactivity batting average.

2018-10-25

2022-05-27

Shows how the Badapple formula for compound scaffold promiscuity combines weight of evidence with bioactivity batting average.

2020-07-14

2022-05-30