Technology is ‘The Other’ with whom Humans Engage the most

The Contemporary Media Environment (CME) is the current situation of the widespread connected world of computing, which features the pervasive presence of technology in an increasingly rich information environment between and amongst human and machine entities.

One aspect of the CME is the increasing emergence of technology as ‘the other’ in the human-technology relation. Humans are now in a wholly new conceptualization and interaction with technology, and also information, where non-human entities are the primary other party in the majority of interactions (Floridi 2014). Technology is ‘the other’ with whom humans are engaging the most.

The theme of the ‘technology other’ has often been explored in film, with the increasing trend of humans and technology being portrayed in full partnership, for example in Big Hero 6 (2014), Her (2013), and Robot & Frank (2012).

Another way that the CME is manifesting the technology other is through embodiment, and in an escalation in the forms and types of human interaction. The technology other is no longer conceived narrowly as Amazon and Netflix recommendations, but instead as a fully-embodied agent. An example of this is robotic personal assistants for home and work like Robotbase’s Personal Robot, MIT’s JIBO, and Amazon’s Echo. Likewise artificial companions, for a variety of functional interaction with humans, may be the next innovation.

A sense of embodiment might also be perceived with advanced voice assistants like Apple’s Siri, Google Now, and Microsoft’s Cortana; they are a new kind of object-person.

Even beyond technology-as-other is technology-as-partner: the best ‘worker’ for many contemporary jobs in the automation economy, perhaps soon to be the machine economy, is a human and a machine in collaboration (Cowen 2013, Carr 2014).

iSchools: Contemporary Information Technology Theory Studies

The perfect merger of academic rigor and contemporary thinking has come together in the concept of iSchools, which give practical consideration and interesting learning opportunities to the most relevant issue of our time: information. So far there are over 50 worldwide iSchools; a global pool, like bitcoin for academia. The March 2014 conference was held in Berlin and the March 2015 conference will be at UC Irvine. With higher education under reinvention pressure from all directions, the possibility of making institutional learning relevant again cannot be underscored enough.

iSchools are the perfect venue to take up not just the practical agenda within the information technology field but also the theoretical, philosophical, and societal dimensions of the impact of information technology. There have started to be some conferences regarding ‘big data theory’ (Theory of Big Data, University College London, Jan 2015), and a calling out of the need for ‘big data theory’ (Big Data Needs a Big Theory to Go with It, Scientific American, Rise of Big Data underscores need for theory, Science News). These efforts are good, but mostly concern having theory to explain the internal operations of the field, not its greater societal and philosophical effect. In addition to how ‘big data theory’ is currently being conceptualized, an explicit consideration of the general theoretical and social impact of information technology is needed. Floridi’s distinction re: philosophy of information is apt; the main focus is how the field changes society, not the internecine methods of the field.

Research Agenda:
Contemporary Information Technology Theory Studies 
Here is a thumbnail sketch of a research agenda for Contemporary Information Technology Theory Studies. Early examples of topics taken up at institutes and think tanks (like Data&Society) are a good start and should be expanded and included in the academic setting. A more appropriately robust agenda will consider the broad theoretical, social, and philosophical impact of the classes of information technology below that are dramatically reshaping the world, including specifically how our ideas of self and world, and future possibilities are changing.

• Big Data (big data, smart data, personal data, group data)  
• Blockchain Technology (Institute for Blockchain Studies) 
• The Brain as an Information Technology (computational neuroscience, connectome, in-silico simulation, cognitive enhancement) 
• Smartcity, Connected World, Self-driving Vehicles 

Neural Data Privacy Rights

A worry that is not yet on the scientific or cultural agenda is neural data privacy rights. Not even biometric data privacy rights (beyond genomics) are in purview yet which is surprising given the personal data streams that are amassing from wearable computing, Internet-of-Things biosensors, and quantified self-tracking activities. Neural data privacy rights is the notion of considering the privacy and security issues regarding personalized data flows that arise from the brain.

There are several reasons why neural data privacy rights could become an important concern. First, personalized health data is already a contentious personal data issue, and anything regarding the mind, and mental performance and potential pathology has even more sensitivity and taboo attached to it.

Second, neural data privacy rights could be an issue because it is not difficult to measure some level of the electrical and other activity of the brain, and ever-ratcheting price-performance technology improvements could make it possible to capture and process the neural activity of vast numbers of people simultaneously in real-time. There are already many consumer-available devices that measure neural activity such as EEGs, PPGs, and tMS systems, augmented headsets like Google Glass, Oculus Rift, and foc.us, and other emotion and cognitive state analysis applications using eye-tracking, mental state identification, and affect analysis. 

Does Google Glass come with a Faraday cage?

Third, at some point, big data machine learning algorithms may be able to establish the validity and utility of neural data with correlation to a variety of human health and physical and mental performance states.

Fourth, despite the sensitivity of neural data streams, like any other form of personal data (where two data elements start to constitute an identification), privacy, security, and anonymity may be practically impossible. At worst, there could be malicious hacking, viruses, and spam targeting neural data streams.

Detailed Essay: "Neural Data Privacy Rights: An Invitation For Progress In The Guise Of An Approaching Worry"

What is Big Data and when will it be Smart Data?

Big data is cell phone users having an average of 100 interactions with their phone per day, all of which generate computerized records (100s of trillions of records). Big data is every financial market transaction, every passenger on every airplane flight, every shipped container, every transportation conveyance, every tweet, and every Internet post (all in the 100s of billions or trillions of records). Every transaction for all time.

One area of long-standing data interest is mortgage statistics since mis-estimating prepayments can cost investors billions of dollars. This raises the question of how prepayment risk is still being mis-estimated. Irrespective, mortgage data is one of the fastest growing kinds of data, both by row and column of tracked data, growing at more than 2x Moore’s law on a log chart (Moore’s law reflects the hardware on which the data is stored and manipulated (algorithms somewhat fill the gap)). This begs the question of smart data rather than big data.

There is much talk about all types of data growing (and data scientists being the biggest category of job growth), but the size of big data should surely be one of its most basic attributes. What is much more relevant is the value that big data provides through its use. For example, how has having more rows and columns in mortgage-tracking spreadsheets improved (if at all) prepayment prediction?

Like genomics, many big data problems are in the early stages of ‘the diffs,’ not knowing which part of the data is salient to keep out of the 99% that may be useless. ‘The diffs’ are the differences, the differences between a sample data set and the reference/normal data set that constitute salience and allow the rest of the data to be discarded.

Big Data: Reconfiguring and Empowering the Human-Data Relation

A strong new presence in contemporary life is big data (the collection and use of personal data by large institutions). As individuals, we can feel powerless in our relation with data.

At present, the human-data relation is one of fear, distance, powerlessness, lack of recourse, and diminished agency. There is an asymmetry of touch in the human-data relation where data can see and touch us without our noticing or being able to touch back. What is missing from the human-data relation is the capacity for humans to touch data in a meaningful way. The asymmetry of touch leads to an incomplete subjectivation of both the human and the data: big data creates a false composite in trying to model and understand the whole individual from just a few electronically-traceable activities, while humans have almost no sight or conceptualization of the entity that is big data.

There are at least two ways to humanize and improve the human-data relation. One is reconceptualizing subjectivation and personal identity as a malleable and dynamic association of elements and capacities, and the other is reconfiguring the power relation between humans and data. To balance the power relation so that humans are more empowered, non-profit institutions, watchdog organizations, and community groups could be created for the defense of personal data, and privacy could be overhauled as a practical impossibility and recast into a system of access rights and responsibilities conferred upon the data.

Presentation: The Philosophy of Big Data
Video (in French): La reconfiguration de la relation humaine-données par le toucher

The Post-Human Biocitizen

We find ourselves in a world with a frenetic pace of life sciences bio-innovation emanating from institutional science, startups, and community biolabs. New possibilities abound in a wide range of areas including  personal genomics, regenerative medicine, cellular therapies, anti-aging, microfluidic chips, quantified self tracking devices and apps, Google Glass, Google diabetes monitoring contacts, brain fitness training, wearable computing, IoT, and cognitive enhancement techniques.

At a higher level, two main themes emerging from this bio-innovation are:
1) what is happening with ourselves as human subjects
2) what is happening regarding data

The human subject is in the process of evolving into a biocitzen, being at the center of health optimization action-taking with a layer of quantified self-tracking gadgetry as a first line of defense, then a layer of preventive medicine health intermediaries (like genomic counselors) and peer collaborators in health social networks and community biolabs, and finally traditional public health services as final line of defense. 

Data's role is transforming even more quickly than the emerging biocitizen where the possibility of collecting, integrating, and sharing huge volumes of health data streams is now possible and required for the destigmatization of health issues and realization of preventive medicine. There are four main data streams to integrate: all of the omics (e.g.; genomics, metaboliomics, etc.), traditional health, quantified self-tracking gadgetry (wearables), and personal internet-of-things (e.g.; smart car, smart home). There is an important need to extend the concept of privacy and rethink the attendant rights and responsibilities of data regimes, quick likely with the advent of protective data intermediary services.

YouTube Video: The Post-Human Biocitizen

Presentation (en français): The Post-Human Biocitizen
Video (en français): Les personnes futures comme biocitoyens

State of Tech: Image Corpus Corralling and API Integration

Two of the biggest current tech trends are image corpus ‘corralling’ and API integration. 1 billion images are taken per day and 6 billion photos are shared monthly.

Uploaded photo databases are the new 'data' corpora.

The first moment this shift became clear in Google’s announcement in June 2012 of the ability to recognize images of cats (the most frequently appearing entity in YouTube videos), and in the big data industry’s continual innovation to manage unstructured data like photos. Now, the sheer volume of image-related activity and opportunity for different consumer and commercial applications is making image classification a focal area for the tech industry.

Functionality is being developed for classifying and accessing content – both images and all web content - with tools such as Imagga as a cloud-based image classification program, and OpenCalais, a standard for text-based semantic content analysis and organization. What we might now start to call the corpus characterization ecosystem is expanding into related tools like DocumentCloud that runs uploaded documents through OpenCalais as a service to extract information about the nouns (e.g.; people, places, and organizations) mentioned.

API integration remains an ongoing trend of the year, with integrated API and developer management platforms like Mashery continuing to grow. API integration platforms give companies a means of facilitating and encouraging external developers to access their content to make apps, and give developers a standardized means of accessing large varieties of data content from different sources to integrate in creating a new generation of sophisticated apps and web services. 

Extreme Data shapes Future Cities

With over 50% of the planet living in cities as of 2008 growing to an expected 75% by 2050 (when the population is estimated to be 9 billion), seamlessly transitioning to cities-of-the-future should be a key planning goal for every urban area. In some countries like the UK, there are strategic initiatives underway to create Future Cities and Smart Cities that include sponsoring hackathons for citizens to work with open urban data, and in other cases research centers are leading efforts such as the MIT Senseable City Lab using the wireless Internet-of-Things (IOT) to sense the real-time city.

Some of the more familiar recent innovations that are starting to pop-up include smart electricity meters, electric car charging stations, on-demand bicycle transport depots, aspirations for vertical farms, and in public transportation: mobile apps with on-demand schedules, journey-planning, and real-time transport information. As another sign of the times, the Oxford English dictionary added the term Internet-of-things in August 2013.

Extreme Urban Data 
The biggest trend reshaping all aspects of our lives, the Big Data Era, is driving a whole new tier of Future Cities and Smart Cities apps connecting big data, open data, statistical processing, and machine learning to user-friendly apps, web services, and other consumable front-ends. Killer Apps could focus on practical improvements to daily life and resource-use: adaptive lighting, smart waste, pest control, hygiene management, eTolls, transport and traffic management, smart grid, asset tracking, and parking. Killer Apps can also be political – using crowdsourced data and social media scrapings to create tools that are the bottom-up sousveillance antidote to top-down surveillance as envisioned in David Brin’s Transparent Society, for example, companies using social media-sourced data to predict country instability in real-time like Cytora.

Quantified Self Fourth Person Perspective and Self 2.0

Quantified self trackers¹ are having an increasingly intimate relationship with technology and data flow in mediating their experience of reality. Technology effectively opens up a new perspective (as vaunted by Nietzsche), a fourth person perspective – a new and objective view of the self, possibly on the road to creating the overself (self 2.0). An important and radical aspect of quantified self (QS) activity is its inherent linkage of the former binary of quantified and qualified in three important ways:

1) The QS Act Itself 
The very act of QS’ing fundamentally includes both the collection of objective metrics data and the subjective experience of the impact of these data

2) QS’ing the Qualitative 
QS methods are now being applied to the tracking of (formerly objectively inaccessible) qualitative phenomena such as mood (e.g.; tracking qualitative word descriptors or mapping subjective experience onto quantitative scales)

3) Quant-Qual are part of a Larger Phenomenon 
To understand QS’ing is to see that it is part of a larger more complex process in which the quantified data collection and the qualitative experience of the data are nodes in feedback loops for behavior change. Data, information, understanding, and action are constituent parts of the looping process

¹Quantified self activity: the self-tracking of any kind of biological, physical, behavioral, or environmental information, often with a proactive stance towards action

I want my wearables!

The layer of quantified self (QS) gadgetry starting to surround us (but not encircle us in the Heideggerian sense) is driving both the mindset shift (e.g.; disruption in the notion of the identity of the individual) and technical practicalities (e.g.; new health informatics business models and technology tools) required for a next generation of health science and technology innovation.

The adoption of the current cell phone-based QS applications and level-one QS devices (e.g.; Fitbit pedometers, myZeo sleep trackers, and Nike+ and Jawbone UP fitness trackers) could give way to the potentially rapidly arriving era of wearable computing. Mobile phones have been one of the fastest adopted technology platform to date, and the wearable computing platform could have even faster adoption and allow for even more self-quantification.

Big Data Era: Not just More Data but New Kinds of Data

One aspect of 21st century data literacy is realizing that there is not just more data, but also that there are new kinds of data.

There is a significant shift from the model where ‘all data is salient,’ for example, each entry on a calendar is a relevant appointment, to a model of being able to recognize different kinds of data and appropriate actions related to specific data types. The focus level upshifts to the correlation, trend, and anomaly level of big data abstractions rather than on the unitary level of the data flows themselves.

Daily quantified self-tracking data for example may be useful from a longitudinal perspective and might not need to be reviewed unless there is an anomaly. Another example is that the relevant action might be looking for correlations across multiple data streams. There could be potential linkage between coffee consumption, social interaction, and mood per as this Sen.se multiviz project investigates, finding some correlation between social interaction and mood. 

Discussed at greater length in: Swan, M. Sensor Mania! The Internet of Things, Wearable Computing, Objective Metrics, and the Quantified Self 2.0. J Sens Actuator Netw 2012, 1(3), 217-253.

Core 21c Skillset: Data Literacy

A core 21st century skillset is data literacy, meaning the ability to recognize, understand, and manipulate various forms of data. One way is through visualization, using visual techniques to both represent data, and also as an inquiry tool for finding patterns.

Some of the basics of data visualization are being able to distinguish between ordinal (qualitative) and quantitative data, and selecting corresponding plotting techniques. For example, a bar chart may be best for displaying simple quantitative and ordinal values, a scatterplot for multiple quantitative data values, and a shape-based plot chart for multiple ordinal values.

Beyond the basics, the next step is mastering more sophisticated visualization techniques. Some of these build on information visualization pioneer Edward Tufte’s work and include using small multiples (plotting several similar charts to highlight differences in one variable), bullet charts, sparklines, horizon charts, and adding a dynamic element to visualizations.

Behavior change through information availability and financial incentives

The presence of the big data era is increasingly difficult to ignore, but what is its influence exactly? Even if big data results in some beneficial outcomes, are humans being enslaved or empowered?

One example of behavior change through the availability of big data is personal power consumption. Home electricity consumption was reduced when individuals were able to self-monitor and obtain feedback on their resource use. There were 7-10% reductions with data and feedback [1,2], and a 32% reduction with data and feedback plus incentives [3].

It is a subtle line between the coercion of social pressure and other invisible influences and the ability to make a truly self-empowered choice whilst living in society. One answer philosophers have had to this question is that at least awareness of the issue helps. Irrespective of the source of motivations, the fact that responsibility-taking through greater awareness and access to information has been demonstrated could possibly be extended to have a similarly transformative impact on health and other sectors.

References
[1] Wood G, Newborough M. Dynamic energy-consumption indicators for domestic appliances: environment, behaviour and design. Energy and Buildings 2003 Sep;35(8);821-41.
[2] Parker DS. Pilot Evaluation of Energy Savings from Residential Energy Demand Feedback Devices. Solar Energy 2008 Jan; 13.
[3] Petersen JE, Shunturov V, Janda K, Platt G, Weinberger K. Dormitory residents reduce electricity consumption when exposed to real-time visual feedback and incentives. International Journal of Sustainability in Higher Education 2007;8(1);16-33.

Terahertz information compression era

Information compression eras is an important area of futuretech: the progression from analog to digital and the developing friction for the next era.

Analog and digital are modes of modulating information onto the electromagnetic spectrum with increasing efficacy.

The next era could be characterized by the even greater effectiveness of electromagnetic spectrum control, particularly moving to multidimensional attribute modulation. Already DNA is a potential alternative encoding system with four and maybe eight combinations instead of the 1s and 0s of the digital era. Terahertz networking (Clariphy, Aurrion) and data provenance are early guides in the progress to the next node of information compression.

Excerpted from: "Reality: analog, digital, or information compression continuum?"

The big data graph era

With the start of the big data era and the ability to collect, store and render meaningful numerous data points, the cultural outlook of the world is shifting too. Graphs, graphs, graphs. Individuals and communities have a social graph, taste graph, preference graph, affinity graph, attention graph, intention graph, values graph, emotion graph, health graph and more.

Graphing theory is being applied to many new contexts such as social networks, media consumption, nanotechnology fabrication, gaming, and genomic analysis and could be one of the many data analysis techniques applied to any large dataset. VLDS – very large datasets – and moving back into the cloud mean that sophisticated data analysis and artificial intelligence techniques could be an expected feature of websites just like social networking commentary and gaming elements have become today.

Next-gen computing for terabase transfer

The single biggest challenge presently facing humanity is the new era of ICT (information and communication technology) required to advance the progress of science and technology. This constitutes more of a grand challenge than do disease, poverty, climate change, etc. because solutions are not immediately clear, and are likely to be more technical than political in nature. The raw capacity in information processing and transfer is required and also the software to drive these processes at higher levels of abstraction to make the information useable and meaningful. The computing and communications industries have been focused on incremental Moore’s Law extensions rather than new paradigms and do not appear to be cognizant of the current needs of science, and particularly the magnitude.

Computational era of science
One trigger for a new ICT era is the shift in the way that science is conducted. The old trial and error lab experimentation has been supplemented with informatics and computational science for characterizing, modeling, simulating, predicting and designing. Life sciences is the most prominent area of science requiring ICT advances, for a variety of purposes including biological process characterization and simulation. Genomics is possibly the field with the most ICT urgency; genomic data is growing at 10x/year vs. Moore’s Law at 1.5x/year for example, however nearly every field of science has progressed to large data sets and computational models.

Genomics: highest-impact near-term advance

Genomics is making faster progress than any other technology in recent history. Usually the vista from any point on an exponential curve looks flat to the experiencer but not so with consumer genomics, the field is exponentiating from any vantage point. Genomics scientific research and commercialization issues were discussed with excitement at the first-ever consumer genomics conference in Boston, June 9-11, 2009. (A PDF of this blogpost is available here.)

Summary

1. Advent of the whole human genome: Automatic whole human genome sequencing of all individuals could likely be a reality in the next few years
   
2. Medically actionable now: Genetic data is medically actionable now and becoming increasing more so, particularly in routing higher-risk individuals into earlier screening. It is estimated that each individual is in the upper 5% risk tier for at least one chronic disease.
   
3. New ICT era (information and communications technology): Genomic data requires a significant new level of information processing, storage and transfer. One whole human genome can range from 6GB-8TB in terms of the data currently transferred between researchers.
   
4. Social inevitability: Widespread genomic sequencing appears to be inevitable which has great benefits together with social challenges such as revealing non-paternity (10-15% in the U.S.), terminal disease conditions and reproductive issues (e.g.; recessive carrier status).
   
5. Heightened role of the consumer: Consumers will have unprecedented access to health information about themselves and could take a much more active and self-directed role in their health management, more likely responding favorably than being consumed with their ‘incidentalome.’
   

Genetic tests – what is now available

Physician-ordered tests (generally insurance-reimbursed)

• For some time, physicians have been ordering any number of one-off genetic tests for specific conditions such as Cystic Fibrosis, Huntington’s Disease, breast cancer (mutations in the BRCA1 and BRCA2 genes) and other conditions. Physicians can also order any of the below tests for patients.
  

Consumer-ordered tests (no doctor-order required, unreimbursed)

• Single condition tests (DNA Direct, $200-1,000)
  
• SNP Chip risk assessment tests (23andme ($399, down from $1,000), DeCODEme ($985), Navigenics ($2,499))
  
• Whole genome scan (Knome ($99,500)) or whole exome scan (Knome ($24,500)) [The price just dropped from $350,000 to $99,000, but it would still seem silly to purchase now when a few more zeros might drop off within months]
  
• Personal Genome Project (PGP), Harvard Medical School, genome sequencing for free in exchange for open data publishing, now expanding from ten subjects to 100,000
  
• Family planning genetic screening: Counsyl
  
• Mate compatibility analysis based on immune system variation: ScientificMatch, GenePartner (The next obvious component would be including recessive disease carrier status in the back-end matching algorithms of dating services)

Out of work due to technological advance:

elevator operator, stock broker, physician(?)

Current genomic testing issues: validity and utility
Validity
There are differing levels of data validity depending on which chip array and methodology is used to sequence the genomic data. Illumina reports being at two 9s now (e.g.; 99.99% error free; experiencing one error per 1,000 reads) and is hoping to move to four and then six 9s of quality. Sequencing is done at different levels of coverage ranging from 1x to 30x coverage, meaning how many times a sequence is read; 30x coverage is the most accurate and highest industry standard at present.

A few people who have tried multiple DTC (direct-to-consumer) SNP chip offerings have found consistent genotyping data (e.g.; having a ‘CT’ at a certain SNP), but different interpretations in lifetime risk probabilities as different markers are evaluated and rolled up into risk assessments across the companies. The risk of false negatives and false positives abounds.

Direct-to-consumer genomic testing companies:

Heterogeneous breast cancer markers assessed

Sources: Navigenics, DeCODEme, 23andme

Not only do different services map different markers to meta conditions like cardiovascular disease, but the most relevant medical SNPs are often not included in DTC SNP chips, probably due to patent and cost issues. A notable example is Myriad, which owns patents on the breast cancer-related BRCA1 and BRCA2 genes. This has become the focus of a timely lawsuit brought by the ACLU regarding the patentability of natural materials such as genes and industry norms of how genes are licensed for diagnosis and therapy.

Whole human genome sequencing renders the patented-gene issue moot as anyone having access to their raw data could look up their genotypes for particular SNP/rsid numbers such as those corresponding to the BRCA1 and BRCA2 genes. (Knome customers can do this now). There will be a need for interpretation tools appropriately aggregating multiple risk alleles. Fee-based or open source genomic data interpretation tools like the SNPedia’s Promethease report could proliferate.

Utility

People would like to know definitively if they are going to have a disease but aside from monogenic conditions (for example, Muscular Dystrophy, Huntington’s Disease, sickle cell disease and Cystic Fibrosis), most chronic diseases are polygenic and influenced by many factors. The current genetic testing for these conditions does not deliver a simple Yes/No, but rather assesses the lifetime risk probability for an individual and whether the individual is at higher or lower risk than the average.

There is ample room for risk interpretation mechanisms for polygenic conditions to become more sophisticated, right now the practice is a multiplicative technique, taking the risk value for each genotyped allele associated with the condition and multiplying them together; weighting and cluster-evaluation would be obvious refinements that research may support over time.

Genetic variation and disease causality
NHGRI and other GWAS (genome-wide association studies) researchers find that genes, as they have been studied so far, only account for a small percent of explaining disease. However, studies have been preliminary, the 1,000 genomes studied may not be enough for complete understanding, for example, about 35 common diseases have been found to have widely replicated common variants. One next step targeted by the NHGRI is to look at rare variants, low-frequency (e.g.; 1-2%) GWAS variants with intermediate penetrance, to possibly explain a larger percentage of disease causality. Simultaneously, our systemic understanding of biology is slowly improving, it seems that in many disease cases it may not be the gene or genotype, but rather the number of copies of the same gene (CNVs), translocations, inversions, and other problems with gene expression and DNA repair that are responsible for disease.

Knowledge gap
Genomic technology has been moving so fast that at present, most physicians do not have genetic training. The genetics community is the primary party helping to generate, interpret, present and monitor genomic data. Over time, other communities like physicians and genetic counselors (one of the world’s fastest-growing job categories) will hopefully become helpful in interpreting data together with patients. Genetic training is a key target area of CME (continuing medical education), for example the National Coalition for Professional Education in Genetics' "Genetics Education for Health Professionals: What are the Key Messages? How do we deliver them?” (Sep 2009) and Harvard Medical School’s “What the Primary Care Provider needs to know about the Genetic Basic of Adult Medicine” (Oct 2009).

Medical relevancy
That disease has a molecular basis is now undisputed and medicine is slowly shifting to reorganize around this. Presently, 1,400 genes can be tested to inform various clinical decisions, and 225 are deemed clinically significant. 100 new tests are being added annually. In some cases, medical information exists but is not being used, for example a straightforward marker for poor drug metabolizers, CYP2D6. About 10% of Caucasians are poor metabolizers however this is not routinely tested for ahead of time (nor in the DTC SNP chip tests mentioned above) and the same drugs are given to all patients in a trial and error process, sometimes in lower doses (e.g.; warfarin) due to fear of overdosing those for whom it could be harmful.

Another example of medical relevancy in genomic testing is the NHGRI’s GWAS study finding of the first nine genetic risk variants for type 2 diabetes: TCF7L2, IGF2BP2, CDKN2A/B, FTO, CDKAL1, KCNJ11, HHEX/IDE, SLC30A8 and PPARG; particularly the first one, TCF7L2. Higher-risk individuals identified early in life could receive targeted healthcare.

Additive statistical approach
So far, general genomic testing suggests that on average, each patient is in the upper 5% risk tier for at least one chronic disease (e.g.; cancer, cardiovascular disease, myocardial infarction, etc.) and that there is value in understanding genomic risk factors earlier in life. Whole human genome sequencing automatically at birth could mean a lifetime of personally relevant healthcare.

Although genomic tests do not predict polygenic disease definitively, they are medically actionably in taking conventional risk percentages (e.g.; American female lifetime breast cancer risk = 12%; American male lifetime prostate cancer risk = 16%) and layering on the specific genetic risk of the individual to route higher-risk individuals to screening and therapeutics earlier. Several researchers estimate that the earlier identification of higher risk patients could reduce overall healthcare costs by about ~$100,000 per person per condition.

Patient behavior: a key component of medical actionability
Although there is no known cure for Alzheimer’s Disease, and even a firm diagnosis can only be made at autopsy, Boston University’s REVEAL study has shown that people do change their behavior after receiving a positive diagnosis for Alzheimer’s Disease (mainly through purchasing supplements and some increase in exercise). It is also known that mid-life cholesterol levels correlate with Alzheimer’s Disease, so the highly actionable behavior for someone with an APO E4 positive allele could be more closely managing cholesterol intake.

Family history
The role of family history is another important component of disease prevention, diagnosis and management, and there are starting to be helpful web-based tools for consumers to assemble, manage and access family history data such as My Family Health Portrait.

Technology status
Technology advance has been the key enabler of the genomics revolution. The first genome sequencing project, completed in 2003 cost $3b. Now, the cost of genetic sequencing is dropping to the point where a $100 whole human genome may be available in the next few years, in 2010 according to Pacific Biosciences. There are several next-gen sequencing platforms in process now to supercede the current array-based method.

Next-gen sequencing platforms
Next-gen genomic sequencing platforms are generally falling into two categories, those using synthesis (specifically multiplex cyclic sequencing by synthesis) and those not using synthesis. Some of the most interesting next-gen companies using synthesis are Pacific Biosciences, Ion Torrent Systems and RainDance Technologies. Some of the most exciting non-synthesis-based next-gen sequencing companies are Oxford Nanopore Technologies, and NABsys and Halcyon. NABsys and Halcyon are electromagnetically-based rather than optically-based which means they are not dependent on light or fluorescence so the cameras can go much faster, perhaps 10,000 frames per second. Harvard Medical School maintains a nice overview of current and emerging gene sequencing technologies.

Transcriptome, proteome, metabolome, microbiome…
In addition to improving the cost and speed of existing genomic scanning, sequencing advances could open up the way to the eventual characterization of the whole cell and its interactions through the sequencing of the transcriptome, the proteome, the metabolome, the microbiome and other biological features. In the farther future, histone modification sequencing, DNA methylation, acetylation and phosphorylation are other characterization processes of interest that could be included.

Petabyte data era: processing, storage and transfer challenges
The biggest challenge consuming national genomic research labs at present is data processing and network communications. Genomic data is growing at 10x per year (vs. Moore’s Law growing at 1.5x per year). Research labs have problems with data storage, mapping and access, together with intra-site data transfer and external transfer. Shipping terabyte drives via fedex is the best current data transfer method, and at least one lab finds resequencing data cheaper than storing it.

The raw data of the 6b base pair whole human genome is 6GB, not challenging to store, but challenging to work with, it is not like just opening up and manipulating a word document. New data processing algorithms will need to be developed to interact with whole genome data, link it to reference tools and make it searchable and meaningful. Whole businesses can be formed to focus on genomic data curation alone (a second wind for Google?).

Even though the most basic raw data version of the whole human genome is 6GB, the full collection of files in use by researchers for one whole human genome may reach 8TB. The full works may include an intensity file, a BAN file (binary), a SAN file (searchable) and other files with coordinates, variations and other aspects. Part of the challenge is that appropriate data abstractions from the raw sequencing output are not yet known so all the data is kept. There is not yet a good reference model. Apparently, the Archeron X-Prize for genomics (sequencing 100 human genomes within 10 days or less at a maximum cost of $10,000 per genome) remains outstanding not because it cannot be done, but because the results cannot be recapitulated.

Testing inevitability and social implications
It seems quite possible that initial and ongoing whole human genome sequencing (and eventually, on-demand proteome, metabalome, microbiome, etc. sequencing) would be a routine component of everyone’s EMR (electronic medical record) available to both patients and physicians for ongoing predictive, preventive healthcare monitoring. There are some important social implications of widespread whole human genome testing, for example:

Non-paternity
One genetic issue is non-paternity (studies suggest 10-15% is the ongoing rate of non-paternity in the U.S.). In the era of whole human genome sequencing, paternity would be quite easy to trace. One possible impact is that the divorce rate could increase and single mothers could be stratified into lower economic tiers.

Right not to know
Another genetic issue is that of a person’s right not to know about their medical situation. With improving remedies, the right not to know becomes a lot less important. Also it may be quite straightforward for practitioners to deliver healthcare without breaching the patient’s right not to know their genetic information as they do currently. With more actionable treatments, it could become the social norm to know your genetic profile, to learn about potential conditions and work collaboratively with others with similar conditions in attempts to mobilize long-tail medicine, as PatientsLikeMe health social network participants are doing to run their own clinical trials.

Discrimination
GINA, the Genetic Information Nondiscrimination Act of 2008, protects U.S. citizens from discrimination by employers and insurance companies. It is a step in the right direction, but many are not reassured. The law has some holes, such as not covering long-term care providers, and will have to be strengthened via interpretation as real-life cases arise.

DNA Forensics – Gattaca?
In an age of inexpensive genomic testing, the on-demand testing of other people (such as a prospective mate, business partner, supervisor or tenant), as portrayed in the movie Gattaca, could easily occur; one such example provided decisive evidence in a recent divorce case. DNA privacy would become impossible as a practical matter. DNA privacy would become impossible as a practical matter. However, precisely because everyone would be subject to genetic openness and since the present world is not one of scarcity and control as the dystopian Gattaca, it may be that DNA testing and knowledge would not be a substantive issue. Already, several individuals in support of hastened scientific advance and open medicine have open-sourced their genomic data on the SNPedia or via the Personal Genome Project.

Venture capital investment opportunities
There are many exciting potential opportunities for venture capitalists, entrepreneurs and researchers in helping to realize the genomics revolution. The money is already arriving before the physicians as companies, backed by varying degrees of research, seek to monetize genetic risk. The potential demand for personal genomic products and services could be enormous, for example, the marker for weight-loss products is a $40b/year. Here are some potential opportunities:

• Personalized genetic testing, counseling, supplements and other action programs and remedies, for example, Inherent Health’s Weight Management, Heart Health and other tests, and the APO E Gene Diet.
  
• More DTC (direct-to-consumer) genetic testing and interpretation offerings stratified towards differing enduser tiers (e.g.; the aggressive early adopter, the lay person, the Boomer, the Gen Y’er)
  
• A line of genomic testing services to be offered by spas and private clinics; positioned as a luxury item vs. a medical necessity to accelerate adoption
  
• Next-gen sequencing, and next-next-gen sequencing, innovating the technology and the applications to commercialize the technology
  
• Web-based tools for integrating medical records, family history and genomic data, facilitating data collection, entry and access
  
• Genetic literacy products and services for physicians and consumers
  
• Web-based tools to appropriately and dynamically aggregate multiple risk alleles into chronic disease meta conditions such as cancer and cardiovascular disease
  
• Fee-based genomic data interpretation tools like the SNPedia’s Promethease
  
• Data processing algorithms to interact with whole genome data, making it searchable and meaningful with links to external reference databases
  
• Genomic data curation
  
• Cloud computing for genomic data analysis
• Health social networks or other tools for deep longitudinal monitoring over time by consumers/patients of many complex health factors
  

Conclusion
As our molecular understanding of disease progresses and genomic testing continues to decrease in cost and become increasingly medically relevant, adoption could become extremely widespread almost overnight. Physicians could start to see the additive, precise information conferred by genomic testing as a means of improving the care they now deliver, finding themselves initially encouraged and eventually forced into the genomic revolution. Pharmaceutical companies could start to use genomic testing and pharmacogenomics as a means of improving efficacy in drug discovery and delivery, providing some much-needed assistance to their ailing cost models. Consumers could be radically empowered to become curious about and responsible for self-managing their health with automated easy-to-use tools. Genomics as an enhanced approach to healthcare could transform the quality of life worldwide for all humanity.

High resolution information

High quality data represented in helpful and manipulatable ways is making humans smarter and better able to understand and act more quickly and effectively.

It might seem like there is too much information available but in fact most of it is low resolution.

Web 2.0 tools are evolving to facilitate easy execution of the principles graphical display expert Edward Tufte has been exhorting for years:

• Information should be high resolution, presenting the most appropriate data possible in the most readable format possible.
  
• The efficient presentation of information should allow viewers to maximize their reasoning about the information and minimize their time understanding the interface.
  
• People should be able to discover the information and think about it on their own, not be directed in totalitarian ways with the data.
  
• Charts and images should be seamlessly embedded in text.
• Clarification should occur by ADDING information, in the same easy to read high resolution way.
  

1. IBM's Many Eyes
Many Eyes is a free data visualization platform where data sets can be uploaded (exclusively for the public domain) and visualizations generated in a variety of formats including tag clouds, word trees, histograms, bar charts, block histograms and bubble charts.

For example, it took a few minutes to analyze the main discussion topics in the 17,619 messages of the last eight years of activity on the SL4 Artificial Intelligence email list.

Many Eyes Visualization

2. Swivel
Swivel is a San Francisco-based startup also offering free data visualization services. The platform is less robust than Many Eyes, but the site appears to have over 1 million charts.

Swivel Graph

3. Google Charts API
The Google Charts API allows the seamless embedding of graphical data anywhere in an html page. Google Charts API and other web widgets have been developed to construct Edward Tufte's famous sparklines, and this functionality is being continually improved.

Google Charts API: example, homepage

4. Second Life 3D Real-time Data Streaming Tool
MS Futures Group has built a real-time on-demand 3D stock chart application for Second Life. The application is available open source and could be adapted for the 3D graphical display of any quantitative data.

3D Stock Charts in Second Life

Stratified information

There needs to be an increasing number of levels of information about certain topics, especially technology areas. People want to understand things at different levels and there is tremendous value in receiving information tailored to different tiers. Unless one is in the practitioner area, the full technical detail is not relevant and the salient concepts and issues summary will be most relevant, at a variety of levels of detail. Horizontal studies, like identifying and working with similar patterns and linkages across fields is perhaps as important as the vertical levels.

The need for stratified information is an effect of having so many more areas and sub-areas, especially in science. In the days of Darwin, polymaths could really be technical experts in many if not most fields. While that may still be somewhat possible today, it is challenging and not useful for one to be an expert in one's entire field. For the practitioner, the focus is on the application and extension of the specific. For the generalist, the value is not in being a deep technical expert in multiple areas but rather to be versed in the key concepts and issues of multiple areas.

The need for stratified information is also driven by the sheer volume of information and the limited capability of the human mind to process, comprehend, retain and use this amount and complexity of information. There should be ways for humans to more effectively interact with information, for example to comprehend-on-demand.