Multi-Modality Machine Learning Predicting Parkinson’s Disease

Abstract

Personalized medicine promises individualized disease prediction and treatment. The convergence of machine learning (ML) and available multi-modal data is key moving forward. We build upon previous work to deliver multi-modal predictions of Parkinsons Disease (PD). We performed automated ML on multi-modal data from the Parkinsons Progression Marker Initiative (PPMI). After selecting the best performing algorithm, all PPMI data was used to tune the selected model. The model was validated in the Parkinsons Disease Biomarker Program (PDBP) dataset. Finally, networks were built to identify gene communities specific to PD. Our initial model showed an area under the curve (AUC) of 89.72% for the diagnosis of PD. The tuned model was then tested for validation on external data (PDBP, AUC 85.03%). Optimizing thresholds for classification, increased the diagnosis prediction accuracy (balanced accuracy) and other metrics. Combining data modalities outperforms the single biomarker paradigm. UPSIT was the largest contributing predictor for the classification of PD. The transcriptomic data was used to construct a network of disease-relevant transcripts. We have built a model using an automated ML pipeline to make improved multi-omic predictions of PD. The model developed improves disease risk prediction, a critical step for better assessment of PD risk. We constructed gene expression networks for the next generation of genomics-derived interventions. Our automated ML approach allows complex predictive models to be reproducible and accessible to the community.

Genetic risk factor clustering within and across neurodegenerative diseases

“…Using this SNP set, the merged data underwent munging and additional population substructure adjustment in GenoML. 25,26 Munging consists of pruning provided genotype data for linkage disequilibrium (LD) by removing any highly correlated genotypes in the sample series (r2 > 0.3 within a sliding window of 1Mb as minimum exclusion criteria). LD clumping and pruning was performed at random and was therefore not biased towards associations from any one disease.…”

Mapping the Diverse and Inclusive Future of Parkinson’s Disease Genetics and Its Widespread Impact

“…As we move forward, standardization and harmonization of datasets, as well as automating data processing is key. An example of this is GenoML (https://genoml.com/, accessed on 20 September 2021) which enables automatic machine learning in genetic studies and has been widely applied in the PD genetics field [68,69].…”

A comparative and comprehensive study of prediction of Parkinson’s disease

“…The derived feature sequences of regular and abnormal walking, as well as the three classes A, B, C, D normal, Parkinson gait, Hemiplegic gait, and Neuropathic gait data sets, were compared with the normal data set during the training process. Mozhdehfarahbakhsh et al [63] proposed a Convolutional Neural Network and MRI based deep learning model to predict the Parkinson 's disease and its stages. In Parkinson's disorder always a challenging to identify its stages and its progression.…”