Hardware Reference
In-Depth Information
Chapter 4
Biochemistry Synthesis Under Completion-Time
Uncertainties in Fluidic Operations
4.1
Introduction
Digital microfluidic biochips have emerged in recent years as a promising platform
for implementing laboratory procedures in biochemistry [
1
,
2
]. Biochemical assays,
such as the dilution of samples and reagents, crystallization of protein molecules,
on-chip chemistry for DNA sequencing, multiplexed real-time polymerase chain
reaction (PCR), protein crystallization for drug discovery, and glucose measurement
for blood serum, have been successfully implemented on such biochips.
The precision of fluidic operations is vital for accurately analyzing bioassays.
For example, in the quantitative measurement for glucose concentration in blood
[
1
], accurate measurements cannot be obtained if the mixing time for blood sample
and enzymatic reagent is not precisely controlled. In order to determine appropriate
settings for the completion-times of fluidic operations, bioassays need to be thor-
oughly characterized [
3
-
5
], i.e., fluidic operations must be repeatedly executed and
monitored to obtain statistically significant results about their completion-times [
3
].
Based on these results, a module library is derived to define the completion-
time for each type of operation, and this library is used as the guideline for the
execution of on-chip operations. Table
4.1
shows an example of module library for
dilution/mixing operations [
6
].
However, due to the inherent variability and randomness of biological/chemical
processes [
3
-
8
], the problem of completion-time uncertainties in fluidic operations
still remains after careful characterization of a bioassay. In practical applications,
the completion-times of fluidic operations are random variables [
3
,
7
,
8
]. Therefore,
an oversimplified module library that defines the completion-times of operations as
constants cannot be used to precisely model the fluidic operations. If the module
library is applied as a guideline for the execution of a bioassay, the bioassay yield
may be low due to the uncertainties of completion-time in fluidic operations. Here
the yield of bioassay is defined as the percentage of bioassay instances that can
produce outcome droplets with expected concentrations.
