MoReL: Multi-omics Relational Learning

Authors: Arman Hasanzadeh, Ehsan Hajiramezanali, Nick Duffield, Xiaoning Qian

ICLR 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments on several real-world datasets demonstrate the enhanced performance of Mo Re L in inferring meaningful interactions compared to existing baselines.
Researcher Affiliation Academia Department of Electrical and Computer Engineering, Texas A&M University
Pseudocode Yes The pseudo-code in Algorithm 1 (Appendix A.1) provides the details of the FGW distance calculation procedure.
Open Source Code No The paper does not contain an explicit statement about the availability of source code or a link to a code repository.
Open Datasets Yes We use the same datasets as Bay Re L (Hajiramezanali et al., 2020), i.e. microbiomemetabolite interactions in cystic fibrosis (CF) and gene-drug interactions in precision medicine. Dataset description and graph construction procedure are detailed in Appendix A.2. (Morton et al., 2019; Quinn et al., 2015; Lee et al., 2018; Wagner et al., 2016)
Dataset Splits Yes The remaining 20% of the reported molecules are considered as a validation set and are only used for the early stopping purpose.
Hardware Specification Yes All the experiments are performed on a workstation with a single NVIDIA P100 GPU.
Software Dependencies No The paper states, "We have implemented Mo Re L and all the competing methods in Tensorflow (Abadi et al., 2015)," but it does not specify a version number for TensorFlow or any other software dependencies.
Experiment Setup Yes The temperature for relaxed Bernoulli distribution is set to 0.3. The normalizing parameter γ in equation 6 is 0.9 while α and β in DFGW are set to 1 and 0.5, respectively. We used the exponential decaying learning rate with the decay rate of 0.01 and initial learning rate of 0.01.