Data Tracking Systems - Biomedical Informatics in Translational Research - page 135

designed and developed, how the workflow works, and how each button in the
interface works. The experimental staff understands where the workflow came
from, what data should be supplied when configuring the system, and so forth. The
biomedical informatics or bioinformatics staff understands both fields and should
be able to bridge the gaps. The trainees should function as a team during the train-
ing session so that they can count on each other to get a comprehensive understand-
ing of the whole system. At the same time, no one should stop short of
understanding the essentials she or he is supposed to understand by relying on
others in the team.
It is important that during training system features be well demonstrated and
explained to the trainees so that they can be appropriately utilized. Remember, not
all the trainees are strong in IT, and many of them will not take the time to explore
the features by clicking around the system. (We refer here to a test system, of course,
because blindly clicking around in a production system can be dangerous and thus
should be prohibited.) Good features, if unknown to the end users, may not only be
wasted but also impose extra burden to the users who then have to develop
walk-around steps to make the system work.
7.5.3 Mismatches Between System Features and Real Needs
It is often the case that a commercial data tracking system was designed based on
the needs of the initial customers. Some features might have been developed based
on “reasonable” expectations. Therefore, not all system features match the needs of
subsequent customers. Thus, system customization is inevitable and sometimes crit-
ical. However, when such mismatches occur at the fundamental level of the system,
the problems can be difficult to resolve. For example, one commercial data tracking
system has a built-in feature to request the barcode of a sample vial during an exper-
iment at every step, but in practice many intermediate vials are not barcoded. To
solve the problem, the end users have to punch in arbitrary numbers in the place of
barcodes, and each time these numbers have to be new since the system does not
accept “duplicate barcodes.” Mismatches like this may not only occur in commer-
cial systems but also in in-house developed systems, sometimes due to
miscommunications between the developers and the end users. The solution to this
problem again lies in the flexibility of the system—a data tracking system should
allow easy modification of the workflows.
7.5.4 Protocol Changes and Other Evolutions
Biomedical informatics research is very dynamic, and new technologies emerge all
the time. A protocol may change with time; for example, in gene expression
microarrays, the number of gene probes always increases. The experimental plat-
form may change with time as well due to a variety of reasons. For example, in Feb-
ruary 2007, GE Healthcare made a strategic decision to discontinue the production
of the CodeLink microarray chips, forcing all of its users turn to other experimental
platforms [59].
New experimental platforms need to be supported, for example in genomics,
DNA copy number analysis is gaining momentum [59]. In proteomics, antibody