Back

A Framework for Automated Construction of Heterogeneous Large-Scale Biomedical Knowledge Graphs

Callahan, T. J.; Tripodi, I. J.; Hunter, L. E.; Baumgartner, W. A.

2020-05-02 bioinformatics
10.1101/2020.04.30.071407 bioRxiv
Show abstract

MotivationAlthough knowledge graphs (KGs) are used extensively in biomedical research to model complex phenomena, many KG construction methods remain largely unable to account for the use of different standardized terminologies or vocabularies, are often difficult to use, and perform poorly as the size of the KG increases in scale. We introduce PheKnowLator (Phenotype Knowledge Translator), a novel KG framework and fully automated Python 3 library explicitly designed for optimized construction of semantically-rich, large-scale biomedical KGs. To demonstrate the functionality of the framework, we built and evaluated eight different parameterizations of a large semantic KG of human disease mechanisms. PheKnowLator is available at: https://github.com/callahantiff/PheKnowLator.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

1
Bioinformatics
◐ 1204 papers in training set
Top 0.2%
41.0%
Similar papers in this journal
2
Bioinformatics Advances
◐ 203 papers in training set
Top 0.1%
13.6%
Similar papers in this journal
50% of probability mass above
3
PLOS ONE
● 5266 papers in training set
Top 17%
10.0%
Similar papers in this journal
4
BMC Bioinformatics
○ 457 papers in training set
Top 2%
3.6%
Similar papers in this journal
5
GigaScience
◐ 212 papers in training set
Top 1%
3.3%
Similar papers in this journal
6
Database
◐ 61 papers in training set
Top 0.2%
3.3%
7
PLOS Computational Biology
● 1863 papers in training set
Top 11%
2.5%
8
Scientific Data
○ 209 papers in training set
Top 1%
2.2%
9
BMC Medical Informatics and Decision Making
○ 43 papers in training set
Top 0.9%
2.0%
10
Journal of Molecular Biology
○ 232 papers in training set
Top 3%
1.1%
11
Frontiers in Bioinformatics
○ 49 papers in training set
Top 1%
1.0%
12
JAMIA Open
◐ 42 papers in training set
Top 1%
1.0%
13
NAR Genomics and Bioinformatics
◐ 242 papers in training set
Top 4%
0.9%
14
Journal of the American Medical Informatics Association
◐ 71 papers in training set
Top 2%
0.9%
15
Scientific Reports
○ 3612 papers in training set
Top 75%
0.6%
16
Neuroinformatics
○ 46 papers in training set
Top 0.9%
0.6%
17
Patterns
○ 78 papers in training set
Top 3%
0.6%
18
eLife
● 5828 papers in training set
Top 69%
0.6%
19
Computational and Structural Biotechnology Journal
● 242 papers in training set
Top 8%
0.5%
20
SoftwareX
○ 15 papers in training set
Top 0.4%
0.5%
21
Frontiers in Physiology
○ 106 papers in training set
Top 4%
0.5%
22
Nucleic Acids Research
◐ 1281 papers in training set
Top 16%
0.5%
23
Nature Communications
○ 5641 papers in training set
Top 60%
0.5%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.