Biomedical Engineering Reference
In-Depth Information
protein of choice, either by displacement of a known ligand or by performing multiple
panning experiments on a set of related proteins [62,103].
11.3.2 Decoding of DNA-Encoded Chemical Libraries
As mentioned elsewhere, after panning the DNA tags of the compounds selected
are amplified by PCR. Binding hits are identified by comparing the relative fre-
quency of the individual library member before and after imposing selection pressure
[54-56,62].
During the initial development of DNA-encoded chemical libraries, subcloning
and sequencing of the individual colonies have been used to decode the amplicon
mixtures [60,71,92]. Although Sanger sequencing offers a relatively long and accurate
sequencing read length, it is cost-effective up to 1000 sequences and only useful when
the amplicon population converges rapidly to a small number of active sequences (e.g.,
decoding of library mixtures containing up to 10
2
compounds).
To get high-resolution maps of large DNA-encoded chemical libraries, microarray-
based techniques have been developed. Hybridization of fluorescent DNA amplicons
to complementary DNA microarrays proved to be a versatile strategy, capable of
giving a global profiling of the DNA distribution without being subject to sta-
tistical effects as when sampling individual population members [63,80,92,106].
However, due to physical limitations on the number of probes that can be
arrayed over the support and the low discrimination rate of closely related
DNA sequences, this technology could not be implemented conveniently for the
deconvolution of DNA mixtures containing more than a few thousand different
DNA codes.
Recent advances in ultrahigh-throughput DNA sequencing, which enable the inex-
pensive analysis of more than 10 million sequences per sequencing run [107-111],
provided an invaluable tool for the decoding of last-generation DNA-encoded chem-
ical libraries. Today, using state-of-the-art high-throughput-sequencing platforms, it
is possible to acquire selection information about libraries containing millions of
compounds in less than a few days and at a cost of about 1000 euros per selection.
In 2008, Mannocci et al. described the first implementation of Roche's 454
high-throughput-sequencing technology for the decoding of DNA-encoded chem-
ical libraries [54]. The technology, based on emulsion PCR and subsequent pyro-
sequencing of DNA-coated beads on ultrahigh-density picoliter plate, was initially
able to provide approximately 1 million sequence tags per sequencing experiment
[107]. To date, the latest generation of the 454 GS FLX titanium sequencer yields
up to 3 million sequences with a 400- to 500-bp read length in a run of about 10 h
[108-111].
More recently, an Illumina/Solexa platform [112] has been used in the decoding of
DNA encoded chemical libraries [55,113]. Such technology relies on the attachment
of single-stranded DNA fragments to a flow cell surface. Bridge amplification of the
individual DNA template hybridized on the surface enables sequencing by synthesis
by means of reversible terminators. Although compared to Roche's 454 technology,
Illumina yields a shorter read length (up to 100 bp) and requires a longer run time
