How to get IC50 values from the fluorescence-based assay

Question

[zero_hsy]

Hello:

I want to calculate IC50 for each drug and cell line. The raw data is a lot of fluorescence intensity values (treated/untreated) for each drug and cell line. How to use the raw data to get IC50?

In the paper "Systematic identification of genomic markers of drug sensitivity in cancer cells", It did do it. But I does not fully understand it.

Could anyone tell me how to do it?

Thanks!

[Devon Ryan]

This will depend on what the fluorescent values represent. IC50 is a functional dose, so you need a functional metric. In general, one fits the resulting values with a sigmoidal curve (e.g., using the Hill equation) and uses the resulting fit coefficients.

[zero_hsy]

Thanks a lot!

Could you tell me more detailed information? I really know little about it.

For example,like the following data:

Small Mol HMS LINCS ID    SM Name    Alternative Names    Recorded Plate    Recorded Well    Control Type    Cell Name    Cell Facility ID    assayType    assayFormat    FI_10.24 uM    FI_5.12 uM    FI_2.56 uM    FI_1.28 uM    FI_0.64 uM    FI_0.32 uM    FI_0.16 uM    FI_0.08 uM    FI_0.04 uM    FI:Control Ratio_10.24 uM    FI:Control Ratio_5.12 uM    FI:Control Ratio_2.56 uM    FI:Control Ratio_1.28 uM    FI:Control Ratio_0.64 uM    FI:Control Ratio_0.32 uM    FI:Control Ratio_0.16 uM    FI:Control Ratio_0.08 uM    FI:Control Ratio_0.04 uM    Control FI Avg    ControlB02    ControlB12    ControlB13    ControlC02    ControlC12    ControlC13    ControlD02    ControlD12    ControlD13    ControlE02    ControlE12    ControlE13    ControlF02    ControlF12    ControlF13    ControlG02    ControlG12    ControlG13    ControlH02    ControlH12    ControlH13    ControlI02    ControlI12    ControlI13    ControlJ02    ControlJ12    ControlJ13    ControlK02    ControlK12    ControlK13    ControlL02    ControlL12    ControlL13    ControlM02    ControlM12    ControlM13    ControlN02    ControlN12    ControlN13    ControlO02    ControlO12    ControlO13    MGH Plate Barcode
10062-101-1    GSK1070916    KIN001-216    None    None    None    CAL-51    50008    syto60    adherent    23629    80293    157072    98512    157547    130689    139750    149948    152107    0.06    0.2    0.39    0.25    0.39    0.33    0.35    0.38    0.38    399151.5    437073    378031    386081    407008    361895    369178    415657    382038    389747    411968    403108    152890    418036    455191    425828    416498    427590    397843    384505    395669    465916    411288    429626    415807    408205    415608    421808    395272    369170    404895    406455    389313    415730    384102    380445    353581    421976    400287    407942    453036    407830    390235    019_450_0010834

It would be so nice of you to tell me more about it.

[Devon Ryan]

Not knowing the details of how the experiment is set up I can't go into too many details. In general, to get an IC50 value we want to look at a plot of some sort of functional readout (presumably that's what the fluorescence represents) vs. dose of a compound. The plot generally looks sigmoidal (i.e., sort of S or Z shaped, depending on how things are oriented), with the half-way point between the maximal and minimal values being the IC50. To then empirically derive this value, we fit a curve to the data (e.g., in matlab, or Igor Pro, or R, etc.), ideally using a function where IC50 is one of the curves coefficients. One of my preferred functions is the Hill equation, partly because each parameter has direct biological meaning.

[zero_hsy]

Thanks a lot!

Are you familiar with R? I try to use ic50 package in R.

When I try to use the ic50 package in R, I look at the example in the ic50 package. The code is as the followings:

library(ic50)
data(A549_1,A549_2,Calu1_1,Calu1_2,H322_1,H322_2,HCC2429_1,HCC2429_2)
dir.create("NSCLC_screen")
write.table(A549_1,file="NSCLC_screen/A549_1.txt",row.names=FALSE,col.names=FALSE,sep="\t")
write.table(A549_2,file="NSCLC_screen/A549_2.txt",row.names=FALSE,col.names=FALSE,sep="\t")
write.table(Calu1_1,file="NSCLC_screen/Calu1_1.txt",row.names=FALSE,col.names=FALSE,sep="\t")
write.table(Calu1_2,file="NSCLC_screen/Calu1_2.txt",row.names=FALSE,col.names=FALSE,sep="\t")
write.table(H322_1,file="NSCLC_screen/H322_1.txt",row.names=FALSE,col.names=FALSE,sep="\t")
write.table(H322_2,file="NSCLC_screen/H322_2.txt",row.names=FALSE,col.names=FALSE,sep="\t")
write.table(HCC2429_1,file="NSCLC_screen/HCC2429_1.txt",row.names=FALSE,col.names=FALSE,sep="\t")
write.table(HCC2429_2,file="NSCLC_screen/HCC2429_2.txt",row.names=FALSE,col.names=FALSE,sep="\t")

When I open one of the .txt, there are a lot of values. I guess it is the value of the FI, but I am not sure. Would you know what does the value mean? It represents what?

[Devon Ryan]

I'm not familiar with that package, when I was dealing with IC50 and EC50 values it was in a different context (small molecule interaction with ion channels). The files you mentioned seem to represent 384 well plates. According to the documentation, those are results from a plate reader, so presumably they're fluourescence measures of some sort. See the package documentation first for further questions.