Decision tree learning in Neo4j on homogeneous and unconnected graph nodes from biological and clinical datasets

Mondal, Rahul; Do, Dung M.; Ahmed, Nasim Uddin; Walke, Daniel; Micheel, Daniel; Broneske, David; Saake, Gunter; Heyer, Robert

doi:10.1186/s12911-023-02112-8

Cited by 2 publications

(1 citation statement)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, web technologies offer data management platforms for storage, querying, retrieval and sharing of heterogeneous sets of linked biological data objects along with their semantics in the form of knowledge graphs, as semantic webs. For instance, Neo4j (Mondal, Do et al 2022), AgroLD (Venkatesan, Tagny Ngompé et al 2018) and KnetMiner (Hassani-Pak 2017) provide data management platforms supporting the Resource Description Format (RDF) data model, processed with SPARQL as query language, and RDF schema and OWL standards for storing ontologies and vocabularies.…”

Section: Knowledge Representationmentioning

confidence: 99%

Knowledge-driven approaches to improve genomic prediction in plants

Farooq

View full text Add to dashboard Cite

Many studies have demonstrated the utility of machine learning (ML) methods for genomic prediction (GP) of various plant traits, but it is still largely unclear if and when ML should be chosen over conventionally used, often simpler parametric methods,. Predictive performance of GP models might depend on a plethora of factors including sample size, number of markers, population structure and genetic architecture. Here, we investigate which problem and dataset characteristics are related to good performance of ML methods for genomic prediction. We compare the predictive performance of two frequently used ensemble ML methods (Random Forest and Extreme Gradient Boosting) with parametric methods including genomic best linear unbiased prediction (GBLUP), reproducing kernel Hilbert space regression (RKHS), BayesA and BayesB. To explore problem characteristics, we use simulated and real plant traits under different genetic complexity levels determined by the number of Quantitative Trait Loci (QTLs), heritability (ℎ � and ℎ � � ), population structure and linkage disequilibrium between causal nucleotides and other SNPs. Decision tree based ensemble ML methods are a reasonable choice for phenotypes with allelic interactions and are comparable to Bayesian methods for additive phenotypes in the case of large effect Quantitative Trait Nucleotides (QTNs). Furthermore, we find that ML methods are susceptible to confounding due to population structure but less sensitive to low linkage disequilibrium than linear parametric methods. Overall, this provides insights into the usefulness of ML in GP as well as guidelines for practitioners.

show abstract

Section: Knowledge Representationmentioning

confidence: 99%