Biomedical Engineering Reference
In-Depth Information
libraries with a high hit rate. As a useful reference, Z-score
3 corresponds to the
activity value expected from less than 0.1% of samples in the inactive population.
Small-molecule compounds have diverse physical and chemical properties. For
any assay, some percentage of compounds interfere with an assay and appear as hits
while having no true effect on the biological target. This phenomenon is independent
of the screening method employed. Although the number and identity of interfering
compounds will vary between the assays, the fact of their existence for any assay is
unaltered. These compounds are called
false positives
. The inverse, when a genuinely
active compound appears inactive in the assay (
false negative
), is rarer, but still pos-
sible. Although diverse in nature and copious in numbers, false-positive compounds
are generally innocuous and are easily identified in the follow-up hit confirmation
studies. Conversely, false negatives are generally unidentifiable in screening data and
therefore usually represent lost opportunities.
Arguably, the major reason for false negatives is compound precipitation in the
assays. Great majority of compounds in traditional screening collections have prefer-
ential solubility in organic solvents; they have limited solubility in aqueous solutions
and precipitate at high enough concentration. For some compounds, precipitation
occurs as early as in the compound storage plates, since hygroscopic DMSO tends to
develop a high water content, decreasing the compound solubility. On the other hand,
the major compound insolubility is observed when compound solutions are diluted
into the intermediate dilution or the assay plates. The precipitation is especially pro-
nounced in the intermediate dilution plates, where the concentration of compounds
is too high for an aqueous solution containing less than 10% DMSO. Utilization of
acoustic dispensers and low-volume pin-tools, commonly used at specialized screen-
ing facilities for compound dispensing, helps to avoid the intermediate dilution step,
significantly reducing compound precipitation.
=
12.2.4 Quality Control of HTS Data
HTS data quality control (QC) relies on the performance of the assay as judged by
assay performance measures. All plates within an HTS run are expected to have
roughly the same S/B, S/N, and Z
-factor values. Systemic errors resulting from
failures of dispensing instruments or employed assays are easily identified through
QC of plate statistics. The data from failed plates are usually discarded requiring
retesting of these plates. As most compounds in small-molecule libraries, with the
exception of a focused or cherry-picked collection, are expected to have no effect on
the target of interest, compound wells could be used for additional data QC. Severe
differences in the hit rate between plates, or a repeated pattern of hits within multiple
plates, serve as warning signs of a potential problem. Luckily, plates with detectable
minor patterns of signal distribution could be recovered through HTS data analysis
using appropriate software packages.
Unfortunately, randomly failed wells are not discernible through plate data anal-
ysis. Nevertheless, advanced analysis of the screening results may be used to help
recover these false negatives. Indeed, most traditional compound libraries, composed
of proprietary scaffolds and commercially available collections, typically contain a
