In this post we will continue explaining some important concepts associated with biosensors (see also our previous post about biosensors), such as the specificity, the limit of detection (LOD), the dynamic range, the regeneration capability or the repeatability; that are not as simple to grasp as it may seem at first sight.
The specificity is the ability of the sensor to only produce a response in the presence of the target molecule. Let us suppose we are developing a fluorescence sensor (see our blog post about fluorescence for more information) for detecting a molecule called A. Ideally, other molecules B, C, D, … will not generate fluorescence at all. In practice, molecules B, C, D will probably produce a response of the sensor, especially if they have similarities with molecule A. However, we will consider the biosensor is properly designed and it is specific if, for the same concentration, the fluorescence response of the sensor in the presence of molecule B, C, D… is negligible in comparison with the response obtained for molecule A (see Figure 1).
Other relevant concept is the limit of detection, known as LOD. It is the lowest concentration of analyte (target molecule) in a sample that can be detected, but not necessarily quantified . According to IUPAC , this concentration must correspond to a signal xL with the following formula:
xL = xbi + k · sbi
where xbi is the mean of the blank (no analyte) measures, sbi is the standard deviation of the blank measures and k is a factor that depends of the desired confidence level. The number of recommended blank measures varies depending on the consulted source, although it is usually 10 or 20. The same happens with k, typically 3 or 3.3, but sometimes set to 4 or 10 with a more conservative criterion .
The problem is that sometimes the LOD finishes by being a theoretical concentration whose real meaning can be put into question. As a result, the limit of quantitation (LOQ) is defined as the lowest concentration of an analyte in a sample that can be determined with acceptable precision and accuracy . In some research works the LOQ is treated as the lowest concentration that can be effectively measured. The discussion could end here, but there are also mathematical formulas for the LOQ, usually analogous to the one for the LOD but with a higher value of k (k = 10) .
Why are these concepts (LOD, LOQ) important? Some molecules linked with diseases are employed as biomarkers, that is, they are used for patient diagnosis. Let us suppose that for a certain disease our previous molecule A has a clinical threshold of 1 mg/l in plasma. This clinical threshold means that for a concentration below this value the patient is ill and for concentrations above this value the patient is healthy (it could be the other way round). Obviously, the lowest concentration effectively measured by our biosensor (we are going to name it LOD for the sake of simplicity) must be lower than this threshold or it will be useless for distinguishing between a healthy and an ill patient (case of biosensor 1 in Figure 2).
Nevertheless, our biosensor must also detect concentrations above the threshold for the disease in order to correctly discriminate among a healthy and an ill patient. It may seem obvious (if our biosensor detects little concentrations it should detect bigger ones, no?), but it is not, as it is explained in the following paragraphs.
We are going to assume that for a certain range of concentrations of molecule A, the response of our fictitious sensor (if we continue with the previous example, the fluorescence) will rise with the concentration of molecule A. It could be the other way round (more molecules of A, less fluorescence), but we are going to suppose we have this direct relationship to make it easier to understand.
Therefore, when the concentration of A molecules achieves a certain value, the response of the sensor will saturate, that is, no matter how much quantity of molecules of A we add, the fluorescence will not increase. This value needs to be higher than the threshold of 1 mg/l that we have established for our disease. If not, in our example, we will never be able to know if an individual is healthy. In Figure 2, biosensor 2 has an adequate LOD of 0.01 mg/l, but the highest value of molecule A that can be detected is lower than the threshold. Therefore this biosensor is not useful for diagnosing our disease.
The range that covers from the lowest concentration that can be effectively measured to the highest concentration that can be detected without saturating the sensor is known as dynamic range. We usually talk about a dynamic range of ‘x’ decades, where a decade corresponds to each time the concentration of analyte (molecule A in our case) is multiplied per 10.
In our case, biosensor 3 in Figure 2, with a LOD of 0.01 mg/l and a dynamic range from 0.01 mg/l to 100 mg/l (4 decades) for molecule A would be suitable for diagnosing our fictitious disease. The LOD (0.01 mg/l) is lower than the threshold (1 mg/l) but the highest concentration we can detect (100 mg/l) is higher than the threshold.
Other aspects that are interesting in a biosensor are its regeneration and the repeatability. If a biosensor can be regenerated it means that, ideally, the target that has been detected can be removed and the biosensor can be employed again. In practice, the bioreceptor may be altered when removing the detected target and more operations could be required in order to correctly restore the biosensor. The repeatability consists in studying how many times the biosensor can be regenerated and how much its properties (LOD, dynamic range) vary after each regeneration.