Seth Berkley's TED Talk on HIV/Flu Vaccine Development

Seth Berkley's recent TED talk focuses on HIV and influenza vaccine research.  In general, I think the talk does a good job of reviewing why it is so hard to develop an AIDS vaccine or a universal flu vaccine.  However, there are some times when I thought that Dr. Berkley was overselling the research results.

For example, Dr. Berkley says "let's take a look at a video that we're debuting at TED, for the first time, on how an effective HIV vaccine might work."  Now, I think the video does do a very good job illustrating the general principle of how vaccines work, but it does not provide any specific details regarding how an effective HIV vaccine can be developed.

At another point in the video, Dr. Berkley suggests that a universal flu vaccine can be created by designing vaccines that target conserved regions on the surface of the influenza vaccine.  These proteins would be located in roughly the equivalent of the blue region of the following 3D rendition of a flu virus (from http://johnfenzel.typepad.com/john_fenzels_blog/images/flu_image.jpg):

As you might imagine, these proteins have not been used because the scientists believed that the immune system would not respond well to them because the H and N spikes (the green and yellow things in the picture above) would block most antibodies produced during the immune response.  Judging from a quick search of the internet, it seems like most images agree with the picture shown above (and in the TED talk).

To be fair, I found one example of a flu virus with less densely packed surface proteins, and the candidate proteins (M2e proteins) may be large enough to clear enough room to interact with the host antibodies.  However, I fear this new vaccine design may be based on data which shows encouraging results during pre-clinical research but is not very effective during clinical trails.  That said, I would obviously be pleasantly surprised if this design does lead to a successful universal flu vaccine, and I honestly do think Dr. Berkley does a good job of broadly describing of how new technology can aid in rational vaccine design.

I also thought that Dr. Berkley did an excellent job describing how changes in vaccine production could significantly increase the effectiveness of flu vaccines.  Namely, Dr. Berkley points out that flu vaccines have been produced from chicken eggs ever since the 1940s.  Different flu strains vary in their ability to grow in chicken eggs, and production of flu vaccines using chicken eggs takes "more than half a year."  Dr. Berkley proposes a method that would allow companies to produce flu vaccines in E. coli.  I think this is an excellent strategy that could significant improve the process of vaccine development.

I think it is also worth mentioning that Dr. Berkley does acknowledge how hard it is to predict the future of vaccine development.  When asked to give a time line to expect an effective HIV vaccine, Dr. Berkley responds "everybody says it's 10 years, but it's been 10 years every 10 years."  In general, it is always important for people to always interpret preliminary research findings with a grain of salt.

Overall, I think Dr. Berkley does a good job providing an interesting talk about a very important subject.

Should the FDA regulate direct-to-consumer genetic testing?

Last week, Walgreens reversed its decision to provide "spit kits" for Pathway Genomics' genetic tests due to a letter from the FDA requiring Pathway Genomics to either get FDA approval or explain why they are exempt from approval.  This week, CVS also made a similar decision to postpone selling of the Pathway tests.

Recently, I have noticed a number of blog posts that seemed to side with the FDA.  For example, 80beats has criticized the Pathway Genomics tests in terms of usefulness, legality, and unpredictable public response.  Genomeboy has complained that Pathway Genomics is not being very transparent in terms of explaining their analysis (especially in comparison to 23andMe).

I agree that a lack of transparency and usefulness would be serious problems for genetic tests.  As mentioned in my first blog post, there are indeed many problems with the current accuracy of genetic tests (in terms of discrepancies between companies, incorrect prediction of clearly known characteristics such as eye color, etc.).  I think it is also essential that companies provide you with your SNP data so that you can try to seek alternative opinions regarding how to interpret the genetic test.  I was also very disturbed that the CSO of Pathway Genomics claims that "[Pathway Genomics] don't feel that [they are] practicing medicine" even though they wish to sell genetic tests in a drugstores.

That said, I think FDA regulation should only be used as an absolutely last resort.  First, I think many other criticisms are not warranted.  For example, the Wall Street Journal has a nice article about how the negative public response to genetic tests has been exaggerated.  Second, I think it will be valuable for consumers to have access to objective analysis of the accuracy and usefulness of commercially available genetic tests (either conducted through public or private means), but I don't think that necessarily has to be done through FDA regulation.  For example, a large database (perhaps something like PatientsLikeMe) of genetic test results, medical history, and lifestyle changes can help provide the necessary information to consumers.  It will take a significant amount of time to throughly examine these genetic tests, and I think this analysis can be conducted much quicker (and perhaps even better) if the public has direct access to the tests.

I would be ok with some sort of warning label that the tests are not completely accurate and other analysis should taken into consideration when making medical decisions, but I think it would be too extreme to completely remove these tests from the market.

What defines a "good" baby?

Yesterday, I noticed this New York Times video showing that babies would prefer "good" puppets after viewing puppet shows with "good" and "bad" characters.  The conclusion was that infants may already have the ability to determine right and wrong behavior.

I think this study could be interesting in so far as it shows that the babies may have to ability to make abstract connections and infer which individuals are likely to be helpful without directly interacting with them.  For example, it is one thing for the baby to learn that it can get fed if it cries.  It takes an additional cognitive step to be able to guess that somebody is likely to help you without ever being helped by that person in the past.

However, I do not think this study really says anything about the origins of "good" or "bad" behavior in the baby itself.  Having the ability to determine that somebody is friendly does not necessarily mean that you have a conscience or a desire to emulate that behavior.  Con artists can manipulate individuals who have a desire to help other people.  Serial killers have lured their victims into cars by pretending to be weak or injured and asking for assistance.  In other words, I do not think the behavior shown by the babies in the study is a reliable indicator of "good" behavior later in life.

That said, I don't believe that moral behaviors absolutely must be learned from parents or other moral authority figures.  If I recall correctly, there have been other studies with babies that actually require the babies to perform truly altruistic actions.  I'm also certain there are several situations where individuals would be "nice" based on rational decision making (even in the absence of the individual's ability to emphasize with other people).  I just disagree with the conclusions drawn from this specific study.