Due to their complex presentations and the limitations of traditional diagnostic methods, rare genetic
disorders have always been among the most difficult to diagnose. Many conditions remain undocumented
for several years, which has led to delays in both treatment and interventions. The increase in multi-omics
data, including but not limited to genomics, proteomics, metabolomics, and transcriptomics, opens up new
avenues in regards to these challenges by providing a wide look into the biological systems of an
individual. Adding several omics layers together increases the possibility of going towards an accurate
diagnosis; the problem is that this is a limiting factor for the effective use of such complexity. ML now
promises a way out from this complexity. This is made possible by the use of ML capability: processing big,
multi-dimensional data sets to find patterns and correlations that might otherwise have been missed.
Recent breakthroughs in ML, including deep learning and transfer learning, also reflect their potential for
integrating multi-omics data and improving early diagnosis for a rare genetic disorder. Still, this direction
has been poorly represented by research papers, at least with respect to the use of ML in diagnosing a rare
disease. This research will work on formulating an ML framework with the capability to integrate multiomics data for the prediction of rare genetic disorders. The hope here is that, through availing the full
capacity of ML in the management of complex interactions among data, this research may be useful in the
improvement of early diagnosis and treatment of these conditions. Beyond that, the research hopes to
enrich the emerging sciences of personalized medicine for future applications of ML to diagnostics of rare
diseases and beyond.