Recently I enjoyed some poolside time and did some thinking again. The funny thing is that when your mind is no longer directed to the feet of daily work, it starts wandering in all kinds of directions, as if it was not mine. I admit, that I did feed it with the thoughts of Nobel prize laureate Daniel Kahneman. Don’t know if that triggered the apparent self-life. I noticed that sometimes it is worth the adventure to float with it and see where you end up.
More is not Less
If I did understand Daniel correctly, the mind forms a coherent image with bizarre speed and accuracy using the impressions that were collected earlier, and sometimes really longtime ago. Many impressions make the most accurate picture. 10,000 flight hours distinguishes a dilettante from a chess master, who can say in one observation: white – mate in 3.
With that assurance, I guess one’s view and thinking is, even though not always true, to a very large degree coherent. In my case, I was thinking about the use of data in precision medicine and specifically in oncology.
We know that population calling (multi-sample variant detection) provides a means to improve quality, detect and suppress artifacts and show cohort patterns. We also know that somatic calling, which is the art of detecting tumor specific mutations in a heterogeneous sample, is not very consistent between tools. Mutect, Strelka and VarScan make very distinct choices when it comes to somatic calling. We do not yet see that field converging to a common ground, as we see with germline variant calling with GIAB as a frequently used benchmark. So, these tools are either too error sensitive, or too truth insensitive.
We further know that biomarker search is done using cohorts. So, what if we would use the output of somatic calling as input for population calling? Now we can use the power of the group to enhance the call quality in each individual sample. We further have the ability to perform GWAS studies on samples with the same diagnose, or even tumor samples from different diagnoses to find ‘common’ mutations or structures between the samples and filter system artifacts. Every somatic caller has a method with a thought behind it, and such method will most likely generate artifacts.
Survive increasing entropy
With the recent technology advancements in sequencing and data processing, and the push for precision (stratified) medicine, the cohort approach seems the only logical way to make progress. Specifically, because it allows us to update the taxonomy of tumors based on their molecular profile. The latter may also open up the selection of drugs indicated for certain tumor types to be used in others.
When speaking to Jos, our Chief Business Officer, the other day, we explored a similar idea when talking about the process of drug development. Turning this into a full biomarker driven and guided process from early research onwards, could make it much leaner than today. This would be a requirement for the pharma industry to survive the move to precision medicine, where more drugs guided by companion diagnostics, need to hit a smaller more targeted market. As a result, each drug will generate la lower return on investment. The balancing phenomena in this process are that, when applied correctly, not only the chance of a drug hitting the market (approval), but also the likelihood of having an efficacy of more than 80% is higher. When time to market and efficiency are improved as well, I think the odds for the great inflection of medicine happening without having to replace the existing pharma industry with a new one are pretty good.
Thoughts, thoughts, thoughts
As said, these are mere thoughts of a mind seeking coherency in a series of impressions. Some of the thoughts could even be rationalized. In any case, I wish we get to the point where we have enough data to win the battle against cancer and reduce the havoc this creates in individual patient lives and their beloved ones. If we can help with state of the art NGS data processing technology that will be our contribution.
If you allow your mind to float to let it make sense ‘on its own’ of the impressions the world gave you, I am pretty sure many thoughts will pop-up, which help to get a grip on cancer and implement precision medicine … just a thought. Please share yours with me below.