The 2010's are already decade of enormous progress in biological data. From low price genomes to high throughput protein assays, there's a coming deluge of information about the chemical state of our bodies.
But this data is only as valuable as the ability to relate it to the human condition; people don't want data unless it tells them something useful. It's the vague relationship between the chemical knowledge and the non-chemical that's the difficult aspect of fulfilling the promise of personalized medicine.
Traitwise is an attempt to solve the second-half of that problem -- the "phenotype". By "phenotype" we mean all the aspects of the human condition that are not chemical -- the reactions to treatments, the feelings, emotions, actions, attitudes, and the environment, to name a few. The characteristics that constitute the phenotype are subjective and for this reason we believe that only by crowd-sourcing from both the medical community and the population at large can we build a database that can be both comprehensive and well-sampled. We crowd-source not only the answering of questions but also the asking of questions, permitting anyone to test an idea. We feel that everyone, not just the medical community, is qualified to ask a question and that this is especially true for those who suffer from chronic illnesses.
We believe Traitwise can be to health-related surveys what Wikipedia is to encyclopedias -- a massive and continually growing source of communal knowledge. In the same way that Wikipedia is not a replacement for expert knowledge, Traitwise is not a replacement for clinical advice or the scientific method. Rather, Traitwise is a way to extend and augment phenotypic knowledge. In this respect, the Traitwise engine can be thought of as a "hypothesis generating machine" -- it doesn't answer questions rather it poses intriguing possibilities suggested for further testing be that an advanced scientific endeavor or a simple experiment by a patient between two treatment alternatives.
The utility of Traitwise correlations is not dependent on genotyping or other chemical sampling techniques. Rather, the foundational result from the Traitwise surveys are "phenotype to phenotype" correlations. With even a modest corpus of answers it is possible to make interesting correlative statements in the form: "People who report X also tend to report Y" the power of which is demonstrated by Netflix and Amazon recommendation engines. Some of these correlations are novel and interesting while others are obvious and boring. Two actual examples from the Traitwise database: "The age of one's mother is correlated to the age of one's father" is incredibly obvious while "those with a negative outlook on life also report craving salt" is far from obvious and suggests several intriguing hypotheses. The evaluation of these correlations, like the questions themselves, are inherently subjective and thus we crowd-source the evaluation of these findings as we do the questions themselves.
The database depends on participation of the public and for this reason we believe that it is extremely important that our system give back as much as it takes. It is for this reason that we designed the question stream to be a dialog -- giving back the results as part of the experience. Finally, we believe that learning about oneself in comparison to others is an inherently fun activity and we have focused intently on making the experience for the user compelling, informative, and fun.
As biological chemical data of all kinds become cheaper, we see the Traitwise database as the keystone of connecting that data to human traits or illnesses. For example, soon the price of DNA sequencing will fall to where it will become practical to sequence not only the static DNA genome, but also the daily changes in the transcriptome as well as antibody profiles that reveal the state of the immune system. A hypothetical low cost sequencer that could sequence these streams of data connected to the Traitwise database might be able to make statements like: "An antibody signature has recently increased in your urine sample. People who reported elevated levels of this antibody also reported flu-like symptoms within three days." Or, "People who reported elevated levels of this transcript reported lymphoma within six months." The diagnostic potential of such analysis is massively amplified with a huge communal phenotype database to correlate against. Thus, the Traitwise database is a foundational technology required for these potentially game-changing diagnostic tools.
Even further in the future it may become possible not only to diagnose with RNA and antibody information but also to cure with them. And again the Traitwise database can play a pivotal role in such future revolutionary medicine. Consider for example the best doctor there is -- the human immune system. Every day our immune systems produce agents that keep us healthy using antibodies. Like all the other aspects of our bodies, antibodies are coded for with DNA that can be sequenced -- i.e. converted into bits. Unlike genomic DNA however, the antibody codes change from day to day as they adapt to the constant influx of infectious agents. Thus, with the collective knowledge of everyone's immune system, it seems plausible to fight every disease by somehow transferring the knowledge of one individual's immune system to another person. If this were possible it would mean that for the first time in the history of world it would be possible to pass acquired immunity from one organism to another. The viruses and other pathogens wouldn't know what hit them.
Again, this science-fiction futuristic medicine would require a huge database of phenotype evidence to correlate antibody or transcription state.
It is with these futuristic goals as well as the more immediate goals that we have built Traitwise. We hope that you will support our endeavor by participating.