As scientists, we are only as good as our data.
We collect data using assays; this is a jargony term that essentially means a procedure for making reliable measurements of something. For instance, to assay changes in temperature throughout the year, you may take a reading from a thermometer every day at midday. Now, as it turns out we're not interested in the weather (lucky, considering there are no windows in our lab). We're interested in viruses and we have various assays to measure different things about them.
The quality of our data is determined by the quality of our assays. An important determinant of this is the signal to noise ratio of an assay; that is, how well can we distinguish the thing we want to measure (signal) from the irrelevant background (noise). The graph gives example data collected with one of our assays which measures the ability of hepatitis C virus to get into host cells. The noise here is the measurement made with the negative control. Virus 1 is giving a signal to noise ratio (SNR) of 355 i.e. its measurement is 355 times greater than the negative control. This is good. Virus 2 is giving a SNR of 7, this is not so good. Essentially, this means we are better able to make reliable measurements of virus 1 than virus 2.
Let the games begin.
Recently, we have been paying particular attention to the quality of our data by monitoring the SNR of individual experiments. In fact we've been having a competition: The Signal to Noise Smackdown 2016. For the last six weeks lab members have been posting the SNR for each of their experiments on a scoreboard. The winner was announced yesterday.
Congratulations to Pat Kalemera, MRC PhD rotation student, with an SNR score of 2015!
Thanks to the competition we now have a better idea of how reliable our assays are and we're now looking for ways to improve them, so we can really bring it for The Signal to Noise Smackdown 2017!