High content imaging (HCI) is an increasingly important method for elucidating changes in cellular biology. The combination of validated immuno-fluorescent assays, availability of automated microscopy, and advances in image analysis software has led to increased exploitation of this technology in a variety of settings ranging from cellular signaling pathways to drug discovery. Recent advances in HCI have made the collection of multi-parametric datasets from high-throughput screening routine, but analysis of these data remains a challenge. The few existing analytical tools used to analyze HCI data, usually provided by HCI platform vendors, require extensive operator interaction and, more importantly, lack statistical power. This results in serious under-utilization of the information available from this powerful technology. As HCI applications become increasingly complex, measurements to estimate the composition of the cell population and thus the underlying data structure also becomes more complicated, and the analysis to facilitate the understanding of the biological system will rely heavily on analytical expertise that requires advanced statistical training. The aim of this article is to review the major statistical issues in HCI data analysis, with a focus on quantitative analysis of cell subpopulations.
Get full access to this article
View all access options for this article.
References
1.
AbrahamV.C., TaylorD.L., HaskinsJ.R.2004. High content screening applied to large-scale cell biology. Review. Trends Biotechnol., 22:15–22.
2.
BaldetorpB., DalbergM., HolstU.et al.1989. Statistical evaluation of cell kinetic data from DNA flow cytometry (FCM) by the EM algorithm. Cytometry, 10:695–705.
3.
BlakeR.A.2007. Target validation in drug discovery. Methods Mol. Biol., 356:367–377.
De BrabanderM., De MayJ., JoniauM.et al.1977. Ultrastructural immunocytochemical distribution of tubulin in cultured cells treated with microtubule inhibitors. Cell Biol. Int. Rep., 1:177–183.
6.
DurrO., DuvalF., NicholsA.et al.2007. Robust hit identification by quality assurance and multivariate data analysis of a high content cell-based assay. J. Biomol. Screen., 12:1042–1049.
7.
EudeyT.L.1996. Statistical considerations in DNA flow cytometry. Statist. Sci., 11:320–334.
8.
EvansJ.G., MatsudairaP.2006. Linking microscopy and high content screening in large-scale biomedical research. Methods Mol. Biol., 356:33–38.
9.
GasparriF., CappellaP., GalvaniA.2006. Multiparametric cell cycle analysis by automated microscopy. J. Biomol. Screen., 11:586–598.
10.
GhoshR.N., GroveL., LapetsO.2004. A quantitative cell-based high-content screening assay for the epidermal growth factor receptor-specific activation of mitogen-activated protein kinases. Assay Drug Dev. Technol., 2:473–481.
11.
HastieT., TibshiraniR., FriedmanJ.2001. The Element of Statistical Learning: Data Mining, Inference and Prediction. Springer: New York.
12.
HuangS.2007a. On some basic statistical issues in analyzing high-content imaging data (unpublished manuscript).
13.
HuangS., DowlessM., StancatoL.2007b. Statistical deconvolution of cell populations from high content imaging studies. High Content Analysis, 2007San Francisco.
14.
HuangS., YeoA.A., LiS.D.2007c. Modification of Kolmogorov-Smirnov test for DNA content data analysis through distribution alignment. Assay Drug Dev. Technol., 5:663–671.
15.
HuangP., ChubbS., HertelL.W.et al.1991. Action of 2',2'-difluorodeoxycytidine on DNA synthesis. Cancer Res., 51:6110–6117.
16.
LamparielloF.2000. On the use Kolmogorov-Smirnov statistical test for immunofluorescence histogram comparison. Cytometry, 39:179–188.
17.
LoechelF., BjornS., LindeV.et al.2007. High content translocation assays for pathway profiling. Methods Mol. Biol., 356:401–414.
18.
LooL.H., WuF., AltschulerS.J.2007. Image-based multivariate profiling of drug responses from single cells. Nat. Methods, 4:445–453.
19.
LowJ., HuangS., BlosserW.et al.2008. High content imaging characterization of cell cycle therapeutics through in vitro and in vivo subpopulation analysis. Mol. Cancer Ther., 7:2455–2463.
20.
LowL., HuangS., DowlessM.et al.2007. High-content imaging analysis of the knockdown effects of validated siRNAs and antisense oligonucleotides. J. Biomol. Screen., 12:775–788.
21.
MannR.C., HandR.E.Jr., BraslawskyG.R.1983. Parametric analysis of histograms measured in flow cytometry. Cytometry, 4:75–82.
22.
NicholsA.2007. High content screening as a screening tool in drug discovery. Methods Mol. Biol., 356:379–387.
23.
O'BrienP., HaskinsJ.R.2007. In vitro cytotoxicity assessment. Methods Mol. Biol., 356:415–425.
24.
SchiffP.B., HorwitzS.B.1980. Taxol stabilizes microtubules in mouse fibroblast cells. Proc. Natl. Acad. Sci. U.S.A., 77:1561–1565.
25.
SilvermanB.W.1986. Density Estimation for Statistics and Data Analysis. Chapman and Hall: London.
26.
SpadariS., Pedrali-NoyG., FocherF.et al.1986. DNA polymerases and DNA topoisomerases as targets for the development of anticancer drugs. Anticancer Res., 6:935–940.
27.
StilwellJ.L., GuanY., NeveR.M.et al.2007. Systems biology in cancer research. Methods Mol. Biol, 356:353–365.
28.
TaylorD.L., HaskinsJ.R., GiulianoK.A.2007. High Content Screening: A Powerful Approach to Systems Cell Biology and Drug Discovery. Humana Press: Totowa, NJ.
29.
TaylorD.L., GiulianoK.A.2005. Multiplexed high content screening assays create a systems cell biology approach to drug discovery. Drug Discov. Today Technol., 2:149–154.
30.
TaylorD.L., WooE.S., GiulianoK.A.2001. Real-time molecular and cellular analysis: the new frontier of drug discovery. Curr. Opin. Biotechnol., 12:75–81.
31.
VindelovL.L., ChristensenI.J.1990. A review of techniques and results obtained in one laboratory by an integrated system of methods designed for routing clinical flow cytometric DNA analysis. Cytometry, 11:753–770.
32.
VogtA., LazoJ.S.2007. Discovery of protein kinase phosphatase inhibitors. Methods Mol. Biol., 356:389–400.
33.
WangH., HuangS.2007. Mixture model classification in DNA content analysis. Cytometry A, 71A:716–723.
34.
WilkinsM.R., AppelR.D., Van EykJ.E.et al.2006. Guidelines for the next 10 years of proteomics. Proteomics, 6:4–8.
35.
YoungI.T.1977. Proof without prejudice: use of the Kolmogorov-Smirnov test for the analysis of histograms from flow systems and other sources. J. Histochem. Cytochem., 25:935–941.
36.
YoungD.W., BenderA., HoytJ.et al.2008. Integrating high-content screening and ligand-target prediction to identify mechanism of action. Nat. Chem. Biol., 4:59–68.
