Gene and Tonic: Boxing for Cancer, Dubious Correlations, and When Should a Researcher Retire


In a keynote talk this week for the online Genetics and Genomics conference, computational biology whiz, John Quackenbush, listed some pretty wild correlations found by a Harvard Business School student when he mixed some large data sets.  For example, U.S. spending on science, space, and technology corresponds directly with suicides by hanging, strangulation and suffocation.  We never would have guessed it without the help of big data.

In an L.A. Times article entitled, Why whole genome testing hurts more than it helps, two authors argue that the number of possible connections and patterns in analyzing the three billion bases in the human genome is just "astronomical" and most of the time quite irrelevant to patient care.  They offer more examples of dubious correlations coming from big data seets.  In one study done on 5 million Ontario patients, Canadian statisticians looked for the correlation between astrological signs and hospital diagnoses and found -- are you ready for this? --  "Leos were significantly more likely to be admitted for digestive tract bleeding and Sagittarians were significantly more likely to be admitted for upper-arm fractures.”

Are these spurious correlations the reason that Quackenbush has taken up boxing?  It's the old fashioned methodology:  bang some new relevant correlations into the head.

And finally, we reluctantly come up with several reasons why aging scientists should retire.  There is no mandatory retirement age in the U.S. and younger researchers are understandably feeling neglected by the NIH.  Numbers show that NIH funding awarded to researchers over 65 has doubled since 1998.  And a new plan by the NIH to come up with "emeritus" grants encouraging researchers to wind things down just isn't taking off.  

What to do, what to do?  Enjoy our weekly wrap on life science news.


Customize This

New to Mendelspod?

We advance life science research, connecting people and ideas.
Register here to receive our newsletter.

or skip signup