To provide a review of data, tools, and other resources available for complex and neruodegenrative disorders. Resources available to scientists, parients, caregivers, and physicians.
Neuro Lighthouse is a step to a larger pursuit of healthy aging. There are several open problems and directions which Neuro Lighthouse is trying to address:
- Broadening our perspective of complex disorders: as we have learned, ALS, Alzheimer’s, and Parkinson’s are not single entities, but rather represent a collection of syndromes. There is a broad variability in the clinical manifestations of these complex disorders, which makes the diagnosis, prognosis, counseling, and clinical trial design limited. Instead of analyzing individual disorders, we need to merge multiple disease datasets and let the data-driven methods deconstruct the heterogeneity within multiple cohorts. This will require close col- laboration between areas of medicine, as well as a paradigm shift in our study designs and funding.
- Open science and democratization of tools: one solution to address the health disparities and scientific reproducibility crisis is making data science tools more available and easier to use. For instance, there is a high barrier for junior scientists and non-biostatisticians to work on genomic data. To address this issue and making genomic and machine learning more accessible, we have been working on an automated machine learning tool for genomics called GenoML. There is a need for more similar tools. Making these tools available through upcoming data science platforms such as Terra will revolutionize open science in healthcare.
- Coordination, collaboration, communication, for larger datasets: plethora of worldwide collaborations are necessary to address challenges facing augmentation of the machine learning in healthcare. The challenges include but are not limited to generalizability, diversity, scale, standardization, and comprehensive evaluation.
- Training multi-disciplinary scientists, the CS/ML+Health taskforce: machine learning and computer science are contributing to the clinical understanding of the detection and treat- ment of health and aging-related disorders. However, the lack of a talented and skillful workforce is becoming a major challenge in deploying data science in practice. According to the LinkedIn August 2018 report, the demand for data science skills is rising across in- dustries in the U.S. Consequently, the high demand has resulted in a countrywide shortage of 151,717 professionals with data science skills. This shortage gap is much wider in healthcare, where talent is not only required to have data science expertise but also health- care and biomedical knowledge and expertise. We believe that academia partnered with the government, and the private sector needs to play a national role in training the data science taskforce to empower analytics efforts in health and biomedical science.