Far fetched application for genomics: dating ?

This could sound like a scenario from a sci-fi movie: allowing people to meet and match based on their genetic profiles. However it is reality ; there are already two companies proposing such services : ScientificMatch (US-based) and GenePartner (based in Switzerland).

ScientificMatch claims range the gamut from “better sex life” and are “supported by peer-reviewed articles” (except one) :

1. You’ll love their natural body fragrance–they’ll smell “sexier” than other people.
2. You’ll have a more satisfying sex life.
3. If you’re a woman, you’ll have a higher rate of orgasms.
4. There will be less cheating in your exclusive relationship.
5. As a couple, you’ll be more fertile.
6. Your children will be healthier.

GenePartner bases its matching on a $299 test of HLA, and gives the following explanation :

A greater variety in HLA genes offers a greater variety in possible immune responses. In terms of evolution, this makes perfect sense: children of couples with a higher variety in their HLA genes (and hence, immune responses) will have better protection from a greater variety of diseases.

Only time will tell if these services are (commercially) successful, and even more time to know whether couples formed form their matching live up to their claims. Some have also criticized such services

  1. for lack of scientific soundness (and lack of the ‘magic’ element of a natural match)
  2. for ethical reasons, ranging from privacy concerns to the issue that matching based purely on genes could be assimilated to eugenics.

What is certain is that it relies upon the pervasive feeling that our genetic makeup holds many keys to who we really are, and what life has in store for us. This is probably more true for our prospects for health & disease than for love & romance, but we should expect personal genomics to become more and more visible in our future society.


July 25, 2008 at 3:47 pm 1 comment

InterMedi: Back to collaborative intelligence!

In a previous post, I discussed how the Pubmed API could, in the field of biomedical research, be a useful tool addressing both problems of Pubmed and the crowd of Social Networks. The solution was named “InterMedi: Collaborative Intelligence for Biomed”. This set of tools provided meaningful relations between people, reflecting real life situation. Plus, it offered fun tools such as an equivalent of GoogleFight based both on the number of collaborators and the number of publications.

I promised to work on it within the few weeks after writing this previous post. I did. Now you have InterMedi V2 (though still in beta version).

Old features that are stille available:

  1. Find someone, find his publications, find his previous collaborators
  2. Compare the number of publications and of collaborators of two researchers

New features now available:

  1. Find the set of affiliations of someone’s collaborators
  2. Find a path of collaboration between two persons! (<== this is the greatest novelty)
  3. Easily contact the InterMedi team
  4. Enjoy a new look and feel.

Future developments:

  1. Consulting and posting biomed job offers
  2. Posting relevant biomed ads

So, now, don’t wait a minute and just intelligently collaborate on InterMedi!

July 17, 2008 at 12:24 pm Leave a comment

Collaborative Intelligence for Biomed

This post will not be about personal genomics, and I apologize for this… but still, it will be about personal-something and also still about biomed. I read those two last days, post about ‘Pubmed Fight’ [1] [2]. And it made me smile because I designed such an application previous winter, but slightly different. My tool is a mix between a social network and Pubmed. Let’s see what are the limitations of those:

  • Limitations of Pubmed: while being the most complete database of biomed publications, it has a major drawback: it is just publications-centered, not people centered. Even Hubmed, which is a re-engineering of Pubmed do not cover the people aspect. However the connection between people is an essential element for future collaboration, for knowing who works in the same area as who, etc.
  • Limitations of social networks: Every member of any social network knows it, the “friend” or “contact” status doesn’t always reflect a real-life situation. In my LinkedIn profile or in my Facebook profile, I have “friends” and “contacts” I never saw once in my life. A tool centered only on true relation between people could be welcome.

My tools at interMedi group both aspects. With the “Search for collaborators” applications, you can find who worked with who. With the “Fight” application, you can compare two researchers on the basis of their publication number but also the number of collaborators they’ve had until now.

How does it work?

It is based on the Pubmed API. Based on a writer’s name “A”, you can retrieve all his articles. Based on those articles, you can retrieve all the co-authors. You then have the collaborator’s list of “A”. This tool is in beta version and will/should be developed in the next few months.

By the way, I seize the opportunity to give a link to my personal page.

Pubmed API

This little known API is open, free, and very useful. It is well explained in the NCBI Handbook. I read on another blog post a nice script to get all the chapters of the NCBI Handbook and make just one nice pdf book of them. Read my comments on that post to update the proposed script.

Keep on reading, have fun!

July 4, 2008 at 10:26 am 3 comments

eHealth vanguard

Being focused on personal genomics and its developements through the past few month, it is more than ever good to keep an eye on health data management. Indeed, personal genomics will generate huge amounts of data. It may be useful to realize that, in the same time, some try to build medical data sharing systems. In this article (maybe a bit longer than previous posts), three systems will be compared:

  • US health data management
  • Google Health
  • Belgian electronic health platform: BeHealth

Of those three systems, the belgian one seems the most vanguardist for several reasons that will be covered in the next lines.

1. US health data management

The United States is the only wealthy, industrialized nation that does not ensure universal health care coverage [1]. The US system causes problems to many patients and their providers. They experience problems with care coordination. In fact there is almost no coordination at all. For example, a Harris Interactive survey of California physicians found that:

  • Four of every ten physicians report that their patients have had problems with coordination of their care in the last 12 months.
  • More than 60 percent of doctors report that their patients “sometimes” or “often” experience long wait times for diagnostic tests.
  • Some 20 percent of doctors report having their patients repeat tests because of an inability to locate the results during a scheduled visit.

On another hand, SERMO, an American medical social network, offers its customer (65,000 physicians) to share clinical insights, observations, and review cases. Members must be authenticated as US licensed physicians, based on Social Security Number. Although this system is a good opportunity to share knowledge and even launch clinical trials, this is not really a universal health data management system. Patients have no access to it. Strictly speaking, there is even no “patient record”.

There is thus no integrated, global solution in the USA for health data management.

2. Google Health

Google Health is yet another Google app intended to provide PHR (Patient Health Record). At the opposite of SERMO, it is patient-centered. Even more: the patient is in charge of everything, so that in fact there is no real control on what is true, false, relevant, etc. in the record. The patient is thus responsible of the information his record contains. What interactions are possible with that PHR? Google Health offers to:

  • Search for a medical care provider
  • Import medical records and notices (only from some affiliated US services)
  • Send your information to some affiliated US services

Google Health is not intended to facilitate communication between several specialists or between your physician and some specialist, etc. You have to give the information to the right person(s). A major drawback of Google Health is that the user may feel lost with medical terms. Dealing with medical conditions, lots of names are rather technical. For example, it may be difficult not to get lost in the middle of the 17 types of Anemia listed. The same problem occurs with medications, immunizations, etc. Finally, there is no sign of a potential genomic data integration.

The PHR offered by Google Health is thus a good information management tool (after all, that’s their job) but the user may have a bad experience and feel uneasy with it, being no medical professional. No control about veracity of the information is performed. The integration with other services is really poor. Above all, it is questionable if it is a good thing to send all your medical information to the online advertisement company…

3. Belgian electronic health platform: BeHealth

Belgium has a centralized, universal coverage health care system. Its social security system is one of the most advanced in the world [2]. Every belgian citizen has a social security number and most of the health cares are free of charge.

Family practitioners, hospitals, specialists, social security, all maintain records about parts of the medical history of the patients. Those records are managed by professionals which offers a good quality and relevance of the information. Records are thus not under the patients responsibility but, as a trade-off,  are also fragmented. There is no or few centralization of the patients data. This guarantees privacy but is also an obstacle to more efficient medicine.

A new eHealth project of the Belgian federal government, called BeHealth, is currently being developed. The aims of this platform are:

  • Quality and coordination of health care optimization
  • Patients security optimization
  • Administrative procedure simplification for all actors in the medical field
  • Optimally support health care policy

How will it work? Not by centralizing all the information: this is far too dangerous for privacy. BeHealth will be based on the principle of data exchange between all the medical field actors while guaranteeing the highest standards in terms of security and privacy.

What BeHealth won’t do:

  • Bring modifications to concrete organization of medical cares.
  • Record personal data in a centralized way (what google health does).
  • Monopolize medical electronic services.
  • Realize itself studies and lead experiments on patients data (what sermo does).
  • Take part to political choices or influences.
  • Be technology-driven (what google health is).

This project will allow the routing of appropriate information to medical providers, communicate anonymized data to some (approved) research projects, control the legitimity of access to patients information. More information (in french) here. Since it is based on a collaborative framework, it makes room for every type of medical data, including genomic data. An actor such as GeneTonix (which is a european company with nearly the same objectives as 23AndMe, Navigenics or deCODEme) could fully integrate personal genomic data into BeHealth services.

The Belgian system seems more advanced than the American one. The Belgian project BeHealth is not yet fully implemented. On June 12, Laurette Onkelinks, belgian public health federal minister anounced that a special series of laws to fully integrate BeHealth are going to be discussed. We will follow BeHealth development in the next months with interest.

This article is not exhaustive and it is probable that comparable system already exist in other countries. What about Canada, Australia, …?

July 3, 2008 at 2:56 pm 2 comments

Personal Genome Project

Private companies offering to decode your genome for a price generally reaching thousand dollars exist and become famous nowadays. What is less known, is the Personal Genome Project, which is more a scientific project than a business project. The aim is to get “a critical mass of interested users, tools for obtaining and interpreting genome information, and supportive policy, research, and service communities.” In other words, the PGP offers you to decode your genome for free. There are obviously counterparts to that. Your data will be used by scientists to perform research on it:

We believe individuals from the general public have a vital role to play in making personal genomes useful. We are recruiting volunteers who are willing to share their genome sequence and many types of personal information with the research community and the general public, so that together we will be better able to advance our understanding of genetic and environmental contributions to human traits and to improve our ability to diagnose, treat, and prevent illness.

Although this is an interesting project, with ethical considerations and almost philanthropist purpose, we draw the attention of our readers on several points:

  • First of all, PGP do not guarantee anonymity.
  • PGP not only needs your DNA, but your medical records too and will ask for many details (allergies, immunizations, medical history, medications, physical traits and measurements, diet, ancestry, lifestyle, and environmental exposures). Participants will be asked to periodically update and add to this information.
  • No result is guaranteed
  • PGP will make use of your data for a period of 25 years.

People who want to get their own genome decoded still have the choice: a free decoding solution with no guarantee of results or privacy, or more expensive solutions with at least a high level of privacy and some results. The advantage of PGP still may be that with this 0$ offer, a large cohort of patients could be constituted. And the strength of genomic studies always relies in the number of samples available.

July 1, 2008 at 11:47 am 1 comment

My DNA on my iPod?

If I wanted to store all the information included in my DNA, what size would it occupy? Could I copy it on my iPod and take it everywhere with me?

DNA is a double-stranded molecule. That means it is compounded of two strands, i.e. two sequences of nucleotides (which, once again, are represented with four letters A,C,G and T). Those two strands are complementary, facing each other. A natural measure of the length of DNA is known as “base pair” or “bp” . A base pair is just one nucleotide and its complement on the other strand.

The human genome contains about 3 billions baise pairs. Since there are four nucleotide, if one wants to code it in a computer format, he has to translate it in binary code: with 1’s and 0’s.

For example, let’s say:

  • A=00
  • C=01
  • G=10
  • T=11

Two bits are needed to code four letters. Each letter thus requires two bits of information. Eight bits make a byte, which is a classical measure of computer memory size. Don’t be afraid with the two lines of utterly simple math that lay below… In bytes, the size my genome requires:

3,000,000,000 bp x 2 bits / 8 = 750,000,000 bytes

A Mega byte, or Mb, is 2^20 bytes. In Mb, my genome would need:

750,000,000 / 2^20 = 715.2557 Mb

We can now answer the initial question: YES! I could store my whole genome on any iPod, even the smallest one. However, it couldn’t be stored on a classical CD-R, containing only 700Mb.

For the most tenacious among you, I should precise that here, we made the (sound) hypothesis that we wanted to record only one of the two strands of the DNA molecule. In fact, the two strands are redundant so it is useless to store both of them. But if you wanted to do so, note that the smallest model of iPod (the 1Gb Shuffle) is not able to store your genome!

Be cautious when choosing your next iPod!

June 19, 2008 at 6:02 pm 3 comments

THE genome vs. MY genome

As said in a previous post, the human genome has been sequenced in 2001. Remember the genome is the sequence of nucleotide defining most physical characteristics of an individual. Since then, everyone is talking about “the” human genome. Now a little riddle:

If there is one human genome, how come personal genomics companies now offer to analyze your personal genome?

Is it swindle? Do they sell every time the same information? Of course they don’t. And if you subscribed to their services, you probably paid for valuable information.

In fact the human genome must be seen rather as a map indicating the probabilities of seeing each nucleotide (A, C, G and T) at each location of the genome of an individual. So for example, at a given place in your genome, the nucleotide “A” may be observed. However, maybe at this place the expected nucleotide is “G” in 95% of human beings. This may reveal a specific trait of your genome. 

This is related to SNPs (Single Nucleotide Polymorphism) which surely will be covered in another post. Keep reading! 


June 17, 2008 at 5:24 pm Leave a comment

Older Posts


  • 4,821 hits