High Throughput Assay

What is HTS? (1)

A key activity to generate hit compounds for any defined biological target. Significant investment of time and money. Setting up HTS is often more important than actually running one.

1 of 86

What is HTS? (2)

Screening technique, identify hits, part of drug discovery process (disease, biological target). Identify chemical compounds which can modulate biological target (positive effect, treat disease)

2 of 86

What is HTS? (3)

Express protein, purification, set up bioassay (used in HTS). Try to get reliable hits. Screening large set of compounds (invest time/money). Lay out target in specific format. Prepare solutions of compounds (chemical libraries). Concentration ranges

3 of 86

What is HTS? (4)

Automatic. Screening for days but important to set up

4 of 86

What is the consequence of an assay being invalid or not specific? (1)

Can result in compounds which show activity at one stage but then the hit compounds are not specific enough to be taken to the next stage (go back to the initial HTS, time-consuming process)

5 of 86

What is the consequence of an assay being invalid or not specific? (2)

Specific timeline to manage (competition to get drug to the market first)

6 of 86

How is a bioassay set up for HTS? (1)

A screen of >50,000 compounds or well. Can either be a biochemical or cell based assay. Run in 1536-well plates where possible. Screen between 250k and >>1 million wells. Screening takes from a few days-weeks depending on complexity

7 of 86

How is a bioassay set up for HTS? (2)

Multi-well plate, increase number of wells, reduced volume of assay reagent

8 of 86

Outline the basic HTS process - assay transfer

Identified a biological target, develop assay which shows a response between molecule and target (e.g. change in colour/fluorescence/absorbance/sensitivity), measurable change. HTS platform - screen large number of compounds

9 of 86

How are reproducible results determined?

Carry out manual process and automated process. Compare results for similarity (optimisation on automatic platform)

10 of 86

Outline the basic HTS process - validation

Test set (assays at different concentration ranges, repeats/compare results)

11 of 86

Outline the basic HTS process - review data

Concentration range (determined by strength of interaction between compound and biological target). Determine which concentration gives reproducible results

12 of 86

Outline the basic HTS process - primary screening

Single shot assay (compounds tested once at a single concentration). Need to know parameters (concentration of target in solution, reagents, technique to add concentration of compound). Test against biological target, determine hits

13 of 86

Outline the basic HTS process - confirmation test

Testing depends on hit rate in primary screen. Check results for false positive/negative results. Secondary assay. Work on selective set of compounds (manual/automatic), difference in number of compounds compared to primary screening,

14 of 86

Outline the basic HTS process - XC50s

Determine dose-response (range of concentrations to identify EC50 - concentration at which the compound has 50% activity/binding to target)

15 of 86

Outline the basic HTS process - results database

Compounds can be taken forward, multiple HTS/libraries - drug discovery project

16 of 86

What are the features of HTS workflow?

Bioassay, development. Plate readers (HCS reader - automated microscopy, fluorescence, FRET, FP, luminescence). Cell based assays/sterile/barcode, library management. Automated staining. Lab information management (servers, database, statistics)

17 of 86

How does the HTS reader work for fluorescence?

Excitation of chemical compound, use of particular wavelength, loss of energy (absorption, emission). Lower energy, higher wavelength - fluorescence

18 of 86

What are neutral controls?

Controls in an assay at steady state e.g in an inhibitor assay represent maximal activity in absence of an inhibitor. Consider no inhibitor present, maximal activity at the target

19 of 86

What are scale reference controls?

The activity in the presence of a supra-maximal concentration of a compound, or no added enzyme. Maximum activity of inhibitor possible (increase concentration of inhibitor beyond activity, no increased activity seen)

20 of 86

Describe features of a primary screen (1)

Determines if the HTS was successful or not. Consists of each compound being tested once at a single concentration. Vital for assay to be reliable enough such that a single test is a true representation of the activity of the compound

21 of 86

Describe features of a primary screen (2)

Need to be sure that each plate in the screen is valid and data is comparable across plates and experimental runs. Plates barcoded to ensure there are no missing wells

22 of 86

What are the two types of scenarios for assay results where the cut-off assay is 50%? (1)

Scenario 1 - compound tested twice, results of 60% on day 1 and 35% on day 2. On day 1 the compound would be a hit and on day 2 it would not be selected

23 of 86

What are the two types of scenarios for assay results where the cut-off assay is 50%? (2)

Scenario 2 - compound tested twice, results of 30% on day 1 and 65% on day 2. On day 1 the compound would not be selected and on day 2 it would be a hit. Need to be within 90% confidence interval (not due to chance but method of choice/assay set up)

24 of 86

Describe features of single shot assays (1)

Compounds tested once at a single concentration. Compound activity normalised to 2 controls, neutral and scale reference. Allows companies to compare data across plates/experiments on common scale

25 of 86

Describe features of single shot assays (2)

Separation of control values determine reliability of assay in detecting active compounds in the presence of typical errors of measurements

26 of 86

Describe features of single shot assays (3)

Determine neutral/scale references, want different to be as wide as possible (large window). Expect activity of compound to be in between controls. Close to neutral/not active, close to scale/very active

27 of 86

Describe features of single shot assays (4)

If neutral and scale references are too close - not able to determine reliability, most compound will be in the middle

28 of 86

What are the HTS assay considerations? (1)

Ensure reliability of primary screen (determine which compound to pass onto next process). Number of assays/replicates needed to be confident in assay result (triplicate primary screen, increases workload/resources/costly, balanced decision)

29 of 86

What are the HTS assay considerations? (2)

Single primary screen, duplicated screening or multiple primary screenings and less confirmatory screening. Cost implications. Cost of false negative (could disregard promising compounds)

30 of 86

What are the HTS assay considerations? (3)

Cost of false positive (could pass compound to next process, realise compound is not effective enough, start again e.g. fluorescence polarised assay, coloured compounds, interference with wavelength)

31 of 86

What is a standard for QC of HTS assay data? (1)

Signal: noise ratio (large window between signal and background noise). Larger ratio preferred, test result will fit in between. Window - between SD bars for neutral and scale references. SD of +/-, variation window (results fall within window)

32 of 86

What is a standard for QC of HTS assay data? (2)

Determine extent of inhibition e.g. close to scale reference (close to high % of inhibition), close to neutral reference (low % of inhibition). Assay S/B alone doesn't define quality of an assay and alone can be misleading

33 of 86

What is a better approach for QC of HTS assay data? (1)

Take into account SC around data. Mean neutral control and mean scale reference. A good assay has clear separation between tails of distributions of controls, a large signal window/separation band (>3-fold), transform principle into simple number

34 of 86

What is a better approach for QC of HTS assay data? (2)

Larger separation band, better assay. Qualitative/subjective assessment not preferred. Want to have quantitative assessment (at what separation is good enough to take compounds forward?)

35 of 86

Describe features of the Z'/Z factor - a screening window coefficient (1)

Z' = 1 - (3.SD(N)) + 3.SD (SR))/mean(N) - mean (SR). Z = 1 - (3.SD(Cpd)) + 3.SD (SR))/mean(Cpd) - mean (SR). Small value of Z' if SD is higher. Large value for Z' if SD is lower. Z' of 1 - no SD (ideal assay, get same result each time)

36 of 86

Describe features of the Z'/Z factor - a screening window coefficient (2)

Z'>0.5, less than 1 (some SD expected, shows good assay). Z'>0, less than 0.5 (small separation band), Z'=0 (no screen, no separation band). Z' calculated after running sufficient replicates. Separation band - validation of assay. Z' max value of 1

37 of 86

What are the issues with the Z'/Z factor - a screening window coefficient? (1)

Mean and SD used in calculation are not robust to outliers due to plate position effects etc. Extreme values (outliers) in either neutral/scale reference controls can adversely affect the Z'-factor leading to a poor estimate of Z'

38 of 86

What are the issues with the Z'/Z factor - a screening window coefficient? (2)

Point removal only fixes the Z'factor and is very subjective. Quite often screening platforms chases a Z' value

39 of 86

Describe features of outliers (1)

Extreme values which are not in line with most of the results (not in the trend), results in wide variability. Excluding outliers in biological results would improve distribution (would appear suitable)

40 of 86

Describe features of outliers (2)

But outliers are not excluded (e.g. need to quantify number of patients experiencing S/E, rare, common etc). Z' affected by extreme values. Despite presence of outlier, it doesn't affect quality of data

41 of 86

Describe features of robust statistics (1)

Normal statistics measures work well with normally distributed data with few or no extreme outliers. HTS data contains outliers and whole purpose of a screen is to find outliers

42 of 86

Describe features of robust statistics (2)

In the presence of outliers mean and SD can become inaccurate estimates of centre and spread of data. To ensure good statistical estimates we employ robust measures (they are resistant to outliers and provide good statistical estimates from HTS data)

43 of 86

Describe features of robust statistics (3)

Don't want to eliminate outliers (but don't want outliers to dominate the way the assay results are analysed) - potentially good compounds

44 of 86

Describe features of the median

Central tendency of data (describes data better, not affected by outlier as much as the mean). Mean value could be biased by a single result. Averages can be misleading

45 of 86

Describe features of the robust Z'/Z factor - a better screening window coefficient (1)

Replace mean with the median in the equation. Median and robust SD used in calculation are robust to outliers due to plate position effects. Extreme values (outliers) in either neutral or scale reference controls will no lead to a poor estimate of Z'

46 of 86

Describe features of the robust Z'/Z factor - a better screening window coefficient (2)

No need for point removal (data is what it is), no chasing a Z' value

47 of 86

What is assay validation? (1)

A series of experiments and subsequent data analysis to ensure that the assay is reliably detecting compounds having an effect on the biological target using equipment and reagents to be used in the primary screen

48 of 86

What is assay validation? (2)

In secondary screening, there are known inhibitors used to determine if the assay if reliably detecting active compounds. In HTS there are no known inhibitors. In validating an assay there needs to be replication to assess, reliability of assay

49 of 86

What is assay validation? (3)

Minimum of two experiments across 2 days. Scale reference can be misleading. Use artificial method to produce max response e.g. destroy cells, work with new target, no identified inhibitor

50 of 86

What are the two approaches of assay validation? (1)

Construct a validation set that is representative of the diversity in your screening collection. 1 - for large scale HTS this is ~ 10,000 compounds, select random number of plates from collection. 2 - screen plates at least twice across 2 days

51 of 86

What are the two approaches of assay validation? (2)

Assess level of reproducibility across the repeated data sets, false discovery rate (active once out of 2 replicates), median of differences between 2 replicate values. Need some active compounds to assess validation across activity range

52 of 86

What are the two approaches of assay validation? (3)

(Add in some known inhibitors or construct a set with known frequent hitters)

53 of 86

What do HTS and validation libraries cover?

Chemistry diversity e.g. range of functional groups on compounds - steroids, aspirin, ibuprofen

54 of 86

What are PAINS?

Compounds which have activity at one stage but no activity at the next stage

55 of 86

Which terms are used to define analytic validation of bioassays?

Accuracy, precision, repeatability, intermediate precision, specificity, detection limit, quantitation limit, range

56 of 86

What is accuracy?

How close results are to the known true value, determine spread of data (measure of variation), level claim. Normal distribution of data, may contain outliers (not significant)

57 of 86

A SD of +/- 3 gives what percentage of data on a normal distribution?

99.73%

58 of 86

Describe features of the mean (average) chart (1)

Normal distribution, average in the centre of bell shaped curve. USL and LSL, true data within variation. Abnormal variation (beyond limits). Could calculate Z'/Z, determine whether or not to carry out HTS on compound

59 of 86

Describe features of the mean (average) chart (2)

Preferable to have smaller SD, better Z' value

60 of 86

What is the capable process?

Almost all measurements of a stable process fall inside specification limits. Spread of results can vary e.g. different companies, equipment used

61 of 86

What is precision? (1)

Measure of variation (analogy of dart board for accuracy and precision). Accurate but not precise - reproducibility not suitable. Accurate and precise - replications give similar results. Percent accuracy. Small variation (more precise results)

62 of 86

What is precision? (2)

Precision of analytical procedure is usually expressed as SD or coefficient of variation (RSD). Confidence interval should be reported for each type of precision investigated

63 of 86

What is repeatability?

Expresses precision (spread of data, variability) under the same operating conditions over a short interval of time. Also termed intra-assay precision

64 of 86

What is specificity (selectivity)? (1)

Readout from assay needs to be specific and free from related interferences. Fluorescent polarisation assay or change in fluorescent signal is a common readout. False positive for coloured compounds. Selection of wavelength (free from interference)

65 of 86

What is specificity (selectivity)? (2)

Sensitivity and robustness, sometimes input can vary, output can be clustered (robust process)

66 of 86

What needs to be considered for robustness?

What is the capacity of measurement to remain unaffected by small but deliberate variations in method parameters (how sensitive is the measurement to the controlled variables) - extraction/reaction time, reagents, pH, temperature, flow rate, column

67 of 86

Why may a compound show different activity day to day?

Typical sources of error - adding compound, adding assay reagents, reader errors, edge effects, patterns on plates in general, positioning of controls (can affect variation when results are subtracted from controls), no re-calibration

68 of 86

Describe features of data distribution

In screening, data is not normally distributed. Distribution has long tails, not symmetrical. Long tails mean average/SD are not valid measures. Most compounds near centre (not active), long tail containing active compounds (effects mean, use median)

69 of 86

Describe features of the Bland Altman Plots (1)

Plots the difference between pairs of measurements vs the means of the pairs to assess agreement between data sets. Determine alignment, need two different screenings. Calculate 95% confidence interval. Determine false positive/negative results

70 of 86

Describe features of the Bland Altman Plots (2)

Discard compounds which don't behave as expected. Assess performance of concentration response assay. Identify measure of reproducibility. Drift in assay signal over time (see more +/- in later screening)

71 of 86

Summarise features of data analysis (1)

Does the distribution data look as expected, any systematic plate patterns in the data, comparison of replicates, show a mean difference between replicates of close/zero between experiments. Confidence intervals of average difference

72 of 86

Summarise features of data analysis (2)

Look at both active and inactive compounds (together in Bland Altman plots). Other measures/tests employed but key ones mentions are a good place to start a validation workflow

73 of 86

Describe features of assessing performance of concentration response assays (1)

In HTS compounds only run once in an IC/EC50 screen. Compounds run in other labs (measure of reproducibility is important). Similar QC steps applied as for primary screen - scale/neutral controls, drift assay signal

74 of 86

Describe features of assessing performance of concentration response assays (2)

RZ' not a useful measure in CR assays. Can get meaningful data from CR assay with RZ' of 0. RZ' can be used as an indication of consistent assay performance

75 of 86

Describe features of QC in concentration response assays (1)

Look at variability of repeated IC50/EC50 determinations. 2 ways - borrow information from standard compounds included in every assay (for SAR assays). Repeat number of compounds (30+), assess variability from data, easier to achieve for HTS data

76 of 86

Describe features of QC in concentration response assays (2)

Can assess how close together repeated IC/EC50 determinations are giving rise to measure minimum significant/discriminatory difference

77 of 86

Describe features of QC in concentration response assays (3)

Different results during replications/later screenings (need to re-validate assay). Control pool of compounds. Z'/Z, single shot assay, primary screening, gives idea of spread variability, robustness. Determine IC50, EC50 from dose-response

78 of 86

Describe features of QC in concentration response assays (4)

Z' not a good indication for dose-response. Determine MSD, confirmation screening

79 of 86

Describe features of determining minimum significant difference (1)

Carry out experiment to determine IC/EC50 of 30 or more compounds twice in independent experiments. Select compounds spanning potency range in assay. Calculate pIC/EC50 for all compounds, calculate log potency differences for each compound

80 of 86

Describe features of determining minimum significant difference (2)

2*SD of log potency differences - MSD. Selection of compounds with good activity and some compounds with low activity. pIC50 (use negative logarithm, produce linear scale)

81 of 86

Describe features of reporting HTS results (1)

IC/EC50s reported as molar values on proportional scale. Means errors are not symmetrical. 10 nM +/- 30 nM would mean IC50 ranges from -20 nM to 40 nM on linear scale (cannot have -20 nM). Can only report fold differences +/- 4 fold (2.5 nM-40 nM)

82 of 86

Describe features of reporting HTS results (2)

Report log10 IC/EC50s, put values on linear scale. Report -log10 IC/EC50, get pIC50 and pEC50. As compounds get more potent their pIC/EC50 gets larger. Errors now symmetrical. pIC50 8.0 +/- 0.6 does make sense

83 of 86

Summarise features of HTS in drug discovery (1)

HTS useful in drug discovery. Cannot initiate programme without hit compounds. Replaces traditional bioassay guided drug discovery. Commercial libraries available to screen against specific target

84 of 86

Summarise features of HTS in drug discovery (2)

Specialised screen from CROs (contract research organisation). Automation with robots, expensive, run assay for limited time on automation platforms, want to reduce overhead/cost, screening companies - screen compounds for pharmaceutical companies)

85 of 86

Summarise features of HTS in drug discovery (3)

Offer compound libraries develop bioassay. Combination of HTS and in silico screening (use high performance computing, reduce cost/using certain number of compounds not libraries, reduce time). QC is important

86 of 86

Get Revising

High Throughput Assay

Other cards in this set

Card 2

Front

Back

Card 3

Front

Back

Card 4

Front

Back

Card 5

Front

Back

Comments

Similar Pharmacy resources:

Other cards in this set

Card 2

Front

Back

Card 3

Front

Back

Card 4

Front

Back

Card 5

Front

Back

Comments

Related discussions on The Student Room

Similar Pharmacy resources: