Becoming an Informed Consumer of Personalized Medicine

Patients often put trust in doctors’ to make the right decisions for them.  I recently re-watched Dr. Jon Cohen’s TEDMED talk and heard him accurately describe Americans’ passion and excitement for scalping the best deal on a Samsung or Vizio HDTV.  We know contrast ratios, pros & cons of LED vs LCD, and where to find it at the best price!  But, when it comes to your upcoming medical procedure, do you know your doctor’s complication rate for that procedure?  Are you the guinea pig for their trial run with new surgical equipment?  Dr. Cohen’s thoughts have inspired me to push the envelope further,

I’m going to put even more pressure on American healthcare consumers.  Personalized Medicine (PM) is the buzz word(s) in the healthcare industry today, which should have all of us jumping for joy!  It is bringing new treatment strategies, earlier disease detection and most importantly, risk assessment that we previously lacked.  To Dr. Cohen’s point though, how many consumers really know what personalized medicine is?  It isn’t tangible, you can’t go buy it off the shelf in a pharmacy and it isn’t a “personal” greeter when you walk through the door at the doctors’ office.  In order to understand the concept of PM, one must first understand “Big Data“.


Here’s an opportunity to tackle another intimidating concept that will DOMINATE the future of healthcare.  Pharmaceutical companies (the ones running clinical trials) are getting this right and they will serve as an adequate example of Big Data in action.  Let’s say they have 100 patients from all across America enrolled in a trial.  In order to enroll for a trial, you essentially have to give up an autobiography.  For each of those patients, they record height, weight, hair color, eye color and for some crazy reason…what the patients had for lunch today!  Stage 1 complete, we have all 100 patients with all their clinical data recorded in an excel file or a database.

Let’s also say the Pharma company sequenced two genes for all of the patients (not a small task).  Now we’ve added a lot more excel files to the database.  Here’s where the beauty and simplicity of big data exists.  A scientist performs pattern recognition, also known as analytics, on the data and finds the following:  The 30 people who ate cheeseburgers all have a mutation in Gene A, and the 20 people that ate chicken wings all have a mutation in gene B!  Bazinga!  We’ve found our cause of mutation for two genes (we think).  The bad news is that its more complicated than this simple example.

That example seemed pretty easy, but it wasn’t exactly Big Data.  They only studied 100 patients and only a couple genes.  Real life examples of Big Data in the pharma industry typically include hundreds or thousands of patients all with hundreds of data elements (vitals) and dozens or more genes sequenced.  Completing just one clinical trial is starting to sound like a monumental task, think of all the terabytes of data they are going through, which is why it takes years to complete.  Data elements/vitals that we are concerned with are typically harder to quantify than age, height and weight.  The emerging categories are things like dietary & sleeping habits and exposure to environmental carcinogens like smog or length of time in the sun every day.  The measurements from these categories and gene mutation data are all  wrapped up in many glorified 100,000 row excel tables… for one patient.  Now we’re talking Big Data.  Here’s where the value of bioinformatics shines through.  It’s someone’s job to…. no, no, no, not count the mutations row by row; to write a computer program that will count the mutations, then the program will count similarities across all the patients to try and glean some patterns in the data.  There are thousands of mutations in one person’s genome.  Because there is so much information contained in even one gene, personalized medicine is handcuffed to Big Data.  See my previous post about why it’s important for biologists and doctors to play nice with computer scientists…


We’ve fulfilled part of Dr. Cohen’s call, now we are more “informed” about Big Data, which is the engine that powers PM.  I used gene sequencing as the example above because PM relies heavily on gene sequencing.  Genes provide the instructions for protein production in the human body.  When you have a mutation in your gene it’s be like missing step number 7 when building your IKEA couch.  You just have to guess how to put the couch together and one of three things would happen.  1. You’re smart and the couch turns out fine.  2. The couch will be a little off, slumping in the corner.  3. Silly, you can’t put your couch together without step number 7, it never gets built!  The same thing happens in the human body with mutations.  We acquire mutations occasionally, sometimes they have no effect and sometimes they are deleterious to our proteins.  These mutations can cause cancer, make you have pointy ears, or give you immunity to malaria infection.

Advances in gene sequencing are making our genomes more accessible than ever.  Some of our genes are well annotated, like the BRCA genes for example.  Research has shown that a mutation in the BRCA gene predisposes an individual to breast and ovarian cancer: see Angelina Jolie’s inspiring story in the NY Times.  She went through with that procedure because they found a mutation in her BRCA gene, not because doctor’s found a tumor, not because she was feeling ill.  Whether you agree with her decision or not, you can’t deny the  importance of knowing what information is contained in your genes.

The personalized medicine utopia would be to have every gene sequenced on every human being.  We’re not even close, but we are making strides.  Let’s go back to our simple example about cheeseburgers and chicken wings.  What I didn’t tell you is that those 100 people enrolled in the clinical trial all complained about a rash on their arm.  They ALL were prescribed the SAME drug to clear it up.  Three months later we find out that the drug worked on everybody except the people that ate chicken wings, a mutation in gene B causes drug resistance.  If you extended this approach to real medicine, not my silly examples, some important discoveries can be made by using association mapping from individuals with similar gene signatures.  Because we’re taking the time and effort to record sometimes seemingly unnecessary data, we have the ability to make these associations.


We are poor consumers of “regular” medicine as Dr. Cohen reminded us, however we still have a chance to redeem ourselves.  Studies are published every day about characterizing sub-populations by these things called “snp”s (single nucleotide polymorphisms).  Humans naturally have variation in our genome, whether its your twin sister or some guy from an eastern European country on the metro, your genomes will still be different.  SNPs aren’t exactly mutations, but they do represent a difference in genes from one individual to another.  Because of Big Data and personalized medicine we are beginning to characterize entire populations of snps, which is very exciting.

People are out there getting parts of their genomes sequenced for $100 from companies like 23&me.  Customers get results that tell them what SNPs they have, which is how 23&me figures out your ancestry and crime scene investigators confirm identities.
SNPs represent natural variation in the human genome between unique individuals. SNPs are not mutations.

Certain ethnic populations carry similar SNPs .  Research is constantly being done on snps, papers are published daily that inform people what it means to carry a certain snp.  23&me customers can go see if they have that snp in that particular gene and determine whether or not the research means anything to them.  An example from a couple months ago identified a snp that causes people to have a longer ring finger than middle finger.  Here’s the link to their page outlining newly discovered snps every month.   At this point the results are trivial in the medical realm,  they revolve around ancestry, diet and lifestyle.  However, SNPs are equally as important to understand as genetic mutations as they can have similar consequences.  What if, one day, a study finds that all long ring finger SNP people are good candidates for a drug that cures diabetes.  Hopefully you have that SNP, but if you don’t get your genome sequenced, you may never know.

I’m not saying that everyone should go get there genome sequenced (I haven’t done it yet)!  However, thanks to advances in handling Big Data, research on genes, whole genomes, snps, and mutations are moving at light speed.  So many medical decisions are being made depending on the information in your genes.  It only takes a day to sequence one gene, and another day for your doctor to interpret the results and prescribe treatment.  The FDA is with academia and medical labs step for step on this movement.  Insurance companies are hiring experts to deal with the risk implications of genetic predispositions.  Personalized Medicine is here, it’s a hopeful, inspiring and exciting time to be a consumer of it!

Thanks for reading,


Disclaimer:  To keep with the theme of staying informed, I encourage you to simply google Big Data and Personalized Medicine to get more information.  This is one person’s conceptual approach, I am not the authority on either of these topics.

Image credit:
picture above
Steve Baker, @bakerture,


The New Kid! Bioinformatics as an Emerging Science

image credit:

It’s tough being the new kid.  They often have trouble finding where they fit within the pre-existing social community.  Over time though, their character shines through, reputation grows and they settle into their niche.  The young field of bioinformatics knows the feeling, straddling the line between biology and computer science.  Although most computational biologists (bioinformaticians) have a background in one of the two sciences, they still have trouble “fitting in” with either group.

Venn Diagram shows the overlap of two well established sciences to form a new field of study.
Venn Diagram shows the overlap of two well established sciences to form a new field of study.

Bioinformatics, also known as computational biology, is the study of biology by exploiting the immense computational power that is available today.  It was a science born out of NEED!  There are approximately 37 trillion cells in the human body.  Do we think a biologist sat around and counted those?  Of course not, that would literally take forever.  There are complex mathematical formulas based on existing data that help us arrive at that number.  Bioinformatics’ crowning achievement thus far occurred about a decade ago when the Human Genome Project was completed (April 2003).  The goal, when funded in 1990, was to sequence the entire human genome.  We learned that there are approximately 30,000 genes and 3.16 billion base pairs in one human genome.   The completion of the HGP ignited a curiosity in scientists all over the world and the birth of “Big Data” just occurred in biology.  A graph from the National Center for Biotechnology Information (NCBI) shows the growth of number of genes sequenced over time and the number of whole genomes sequenced (WGS) over time.  Hidden within those sequences are personalized cancer therapies, disease resistance/predisposition and the key to aging.  We just need computational biologists to unlock those secrets.

Why is Bioinformatics under-appreciated in the scientific community?  The answer is two-fold:  fear and the ability to communicate.  Biologists are scared to trust assumptions, computer models and mathematical formulas that computer scientists create.  Also, computer science programmers aren’t trained in biological methods so they do not understand the biology-speak.  This problem left a NEED for people who are trained as both biologists and computer scientists.

Two experiences from my bioinformatics graduate education highlight the problems perfectly.  I studied biology as an undergraduate, so that is where I feel the most comfortable. However, in bioinformatics, you have to be an expert in both biology AND computer science.  Last year, taking an introductory computer science course (deer in the headlights, complete amateur), the topic of  “global variables” came up and I didn’t understand it at first.  I remember asking a graduate student who taught lab section of the course to help explain and he laughed at me, saying that “if I didn’t understand this topic, I was in the wrong program”.  At first, I was discouraged, and I instantly understood why bioinformatics is so hard to “stick”.  As a biologist, you really have to grind through those introductory weeks, months, years… of learning to program.  The second experience made me feel ashamed of biologists.  In the introductory biology course, there were several computer scientists enrolled.  Every week it was a new question about why adenine base paired with thymine or what the difference between the five prime and three prime end of  a DNA strand was.  Stuff I could recite in my sleep.  I felt myself judging them, but quickly remembered the feeling I had when I was in “their” realm, across campus, in the computer lab.  So, I believe patience will be important for bioinformatics to continue to succeed and thrive as a science.

Bioinformatics is here to stay.  I propose a happy marriage moving forward, as science and technology grow together.  Computer scientists should embrace biologists and be willing to teach them the tricks of the trade, not frown at them when they don’t know how to write a block of code “recursively”.  On the other side of that same coin, biologists should show respect and guide computer scientists down the path of good experimental design, rather than pointing out “obvious” flaws in mathematical formulas.  Studying bioinformatics and computational biology has been one of the most rewarding, eye-opening endeavors in my life.  I look forward to the challenges ahead as well as the chance to educate the next generation of computational biologists.

If you’ve made it this far, you’ve read my FIRST BLOG EVER!  Thank you so much for reading, please leave comments, I’m interested to hear what you have to say.

Kevin Arvai

Use the comments link above the article to comment.