Wikipedia defines Personalized Medicine, as a medical model that proposes the customization of healthcare, with decisions and practices being tailored to the individual patient by use of genetic or other information. http://en.wikipedia.org/wiki/Personalized_medicine

However, the attempt to integrate these complex genotypes (Genetic Data) with phenotypes (Clinical Data) is no small task and comes with impediments that I consider to be major gaps in facilitating "personalized medicine". The fours gaps I reviewed below ranges from the need to develop large scale technologies for molecular analysis and clinical genomics studies, the cost of whole-genome sequencing technologies ($1000 genome), the establishment of semantically standards and tools to manage heterogeneous data, and the ongoing need to bridge the gap (the different assumptions) between the biomedical and clinical informatics professionals for the purpose of translating discoveries into practice.

1. Large Scale Technology Challengies

There are large amounts of isolated, heterogeneous and diverse clinical and biomedical data available in literatures and other resources (such as, patient records, lab test results, demographic data, genealogical data, drug-gene association data, genotypic and phenotypic data, pathology reports, epidemiological data) located in silos with no transparent technological means for researchers and clinicians to collaboratively access these data sources to make inferences on genotype to phenotype connections. So the current lack of (development) large-scale technology systems for the purpose of make inferences on genotype to phenotype connections that goes beyond the scope of EHR/EMRs, such that it links patient’s clinical data with genealogical and genomic data to form foundations for translating discoveries into practice is a critical element in the transition towards personalized medicine. I believe that, for us to realize "personalized medicine" the development of large scale multidisciplinary research technology infrastructures are needed to support these types of data integration, from medical records systems, laboratory systems, diagnostic imaging, tissue samples and genealogical records – and the necessary algorithms and tools required for analysis is also needed.

2. Whole-genome Sequencing Technologies Costs ($1000 genome)

The marriage of large scale technologies and molecular genetics analysis towards the study of actual disease is creating advances in every directions and whole genome sequencing is growing as a technology solution to strongly support the transition to personalized medicine. On the other hand the sequencing of protein-coding regions of genomes, instead of whole-genome sequencing is faster, easier and cheaper and there are many literature published on Exome studies on patient with particular diseases, such as cancer. Clearly today, whole genome and Exome sequencing technologies is primarily a laboratory research tool, hence the continued push for further advancement in this field is needed to take this technology into routine clinical environments.  With that said, advancements will come not only from improvement in research and technologies (R&D) but also from the management of disease treatment, informatics skills requirements, the ability to support genome-based EMRs, patient care management and testing cost. From a cost perspective, the dream of the $1,000 genome is just a matter of time, and then disease tissue sequencing for patients would become a routine activity.

3. The Establishment of Semantically Standards and Tools

Genetics and phenotypic information sources consist of both structured and unstructured data that continues to grow rapidly. This unfortunately poses challenges for a diverse data sources to be accessed by researchers, medical research center or bio-pharmaceutical industries for inferences on genotype to phenotype connections. The semantic interoperability challenges between molecular research systems, electronic health systems and knowledge resources are no small matter and call for hours of manual labor. The ability to easily fuse clinical and genomic data, implement the requisite algorithms and systematically accumulate knowledge for further analysis is the common denominator in the personalized medicine equation. So this is a major challenge!

4. Biomedical Researchers and Clinical informatics Professional

Biomedical (BI) Research and Clinical informatics (CI) are two of four primary domains with the scope of Biomedical Informatics. These two professions are closely related but works with different assumptions and objectives; hence the need to bridge the gap between these two disciplines may shorten the time to integrate genomics and phenotypes disease-related (disease-gene) research. While independently these two areas are clearly advancing our understanding of disease, it is important to begin to translate both insights into efforts to translating discoveries into practice under the same roof. CI professionals are instrumental in determining the metabolic activity of disease, patient demographics, matching clinical trials candidates, observation of drug efficacy, and create capabilities for disease-state tracking…etc. However, without the requisite computer science use by BI to aggregate, integrate and interrogate the data needed by CI, many new advances need to generate progress made toward translating discoveries into practice will be limited.

The era of Personalized Medicine to come is no small challenge. It is in need of quality context specific medical knowledge derived from standardized and structured data that can be shared across different professions, clinical/informatics domains and technological boundaries in a secured, compliant manner based upon the rights and responsibilities for the data being accessed. Which leads me to another necessary and critical gap - Regulations (HITECH, HIPAA and IRB...etc). Today's regulations limits the data and information that can be shared across and among research and clinical domains, placing BI, CI and clinicians in a challenging position as they balance obtaining requisite data with patient rights. Additionally, the requisite information systems and science for managing de-identified patient data in support of universal global identifiers will be critical to addressing genotypes/phenotypes integration issues. 

So as you can see, rapid improvements in technology, semantic data structures, informatics professional collaboration and sequencing technologies are not necessarily the only gaps needed for the realization of personalized medicine (improving genomic and phenotypic data integration) but these must be taken into account on how best to exploit the opportunities to facilitate personalized medicine.