hey every little bit helps! Lol
I mine genomic data from clinical trials and look for base pair (DNA) changes (across lots of different clinical trials). So if there's a base pair change that occurs in a..certain number of patients (relative to the total number of patients' data I'm looking at-which is thousands at a time), these can be established as "biomarkers". Then clinicians can test for these biomarkers in other patients that display the disease phenotype, which can lead to diagnosis. Which is 1 area of personalized medicine.
Also, researchers can use biomarkers to "re-purpose" (that's not the word lol) drugs that are already on the market....and do like a virtual clinical trial, instead of a real life clinical trial (less time money effort). To see how a certain population of patients would respond to whatever drug they want to market differently.
I create user friendly pipelines to mine all of this data. I have to create different ones depending on the drug reaction/disease the trial is looking at, cuz you want your program to pick up all the biomarkers that are already known, too. Which is different for different things.
I do this with open source software like QIIME and mothUr, bioinformatics tools like genome reference alignment and variant calling, and languages and packages like R, python, Matlab, and Github.
I also standardize clinical and non clinical trial data in the format that's compliant with the FDA. And they only accept it in 1 format.
It's true that clinical research organizations are paying 23andme millions of dollars to have access to their genomic database (to do virtual clinical trials with). But I guess people had to sign a consent form? Yea I mean I wouldn't count on that if I were you
money talks.
This is 1 branch of data science AND 1 branch of bioinformatics. Which are 2 different things, imo.
And this is all I do now. No more wetlab
(tears of joy lol)