Sapio Whitepaper
Sapio Whitepaper
In fact, the lab environments successful operation is predicated on the successful definition and adherence to
workflows. It could be said that a modern laboratory is an advanced process implementing construct. In making this
statement we are separating the research process that determines what processes will be performed and interpretation
of experimental results as separate from the lab even though these researchers are tightly affiliated with the lab.
Because of the above truisms, the lab is as dependent on workflows as any business unit in any domain today.
Furthermore, the dynamic nature of the laboratory requires agility in the adaptation of old workflows to updated
processes and the quick implementation of new processes as new technologies are adopted by the lab.
What’s interesting is that while the lab is operating as a workflow processing machine, the LIMS software generally has
not been upgraded to support the reality of process driven, dynamic laboratories. This leads to long and costly
implemention timeframes, long learning curves, inflexible software and dissatisfiled end users. It is time to align your
information technology with the essential needs of the modern laboratory.
A Workflow Example
An example of the processes in a genetics core facility might be as follows:
Record information
Unpack the samples and Accession the samples samples including:
Sample arrive at the lab
assess the condition of and print a barcode for •Where it came from
from external locations
the samples each sample •Receipt date/condition
•Phenotypic info, etc
•Draw date
•Sample type
•Shipping method
If you perform genotyping in your lab, it is likely that at a high level the above process is familiar to you. It is be
clear looking at the above that this is indeed a workflow, and in fact, is multiple workflows. We might break down
this overarching process into the following discrete processes:
So the question becomes how this is implemented within a laboratory management software package (LIMS).
Realistically, all software packages allow you to track certain information related to the above processes in some
manner. The question is how this is done. In most cases, the workflow is really a set of instructions on paper or
in some document that says something like:
1. When receiving a new sample, login to the LIMS and go to the sample section. Then click the “Add new
Sample” menu option under the file Menu.
2. Enter the sample ID, then click to submenu for phenotype information entry and enter that information
3. On the main sample screen, find the button labeled “print barcode”, and press it to print the barcode for
the sample.
4. And on and on with menu clicks, button clicks, submenus etc.
The problem with the above is it is a non-linear process that requires the end user to navigate to many different
sections of the LIMS through the life cycle of the samples from receipt to processing to results. In addition, there
is no intelligence built into the process itself. For example, certain steps in a process should not be doable until
certain other steps are completed. This is not reflected in the above process other than via the documentation o f
the process.
Even Microsoft Excel can track data of this nature, but no one would ever call it workflow oriented software
engine. The same is true of LIMS that operate as above; they are not workflow driven but screen\menu driven. A
new approach is needed to mirror the lab environment.
Workflow Requirements
If you were designing a LIMS to operate in this process driven environment, what would you want? Following is a
list for your consideration:
The ability to easily define a workflow with no programming and give it a meaningful name
The ability to define the discrete steps of a workflow (i.e. Tasks) with no programming
Support for many tasks types, such as:
o Tasks that add things to the LIMS for me (e.g. – samples, experiments, projects, plates)
o Tasks that produce barcodes
o Tasks that display a form to the user to collect relevant data for this task
o Tasks that run product extensions
o Tasks that import data
o Tasks that integrate with external systems or lab equipment
Support for a single workflow to work with more than one “item” at a time. Such as 32 samples or 4
plates, or one project, etc.
Support for easily defined task
dependencies with no
programming so the flow
through the workflow can be
controlled (e.g. – step 5
cannot be done until steps 2
and 4 are done)
The workflow should provide
an easy click through step-by-
step process whereby all the
steps are easily visible and
intuitive to the end user. This
means no menu clicking or
extra product knowledge is
really required by the end user
to use the workflow
The ability to mix and match
different records in a single workflow. For example, you might be working with a Project at the beginning
of the workflow, then an Experiment then Samples and then Plates.
Each task should provide easily accessible help for the end use with a single click and without the user
having to exit the workflow
Workflows should be able to be paused and resumed days later
Workflows should allow me to implement business rules that verify data and perform other actions with in
the context of the workflow.
It should be easy to see what steps of the workflow have been completed, and what steps are yet to be
done.
Tasks should be usable across more than one workflow to facilitate reuse.
The ability to export a workflow definition and share it with other labs.
This is a good list of requirements for a workflow system. If these items are in place in your LIMS you have
positioned yourself to implement processes that will mirror your lab processes, reduce training times and increase
throughput. This is not all that matters however; the underlying data model and its flexibility will have a major
impact on the utility and adaptability of both the workflows and the LIMS in general.
The LIMS should provide the ability for the customer to define new data type(s) to be tracked (Data Types may
also be referred to as “Objects”.) Data Types refer to real world (e.g. “Illumina Sample”) or theoretical
(“Project”) items that are to be tracked by the LIMS. With this capability you can track literally anything you
ever want to track. You should not be limited to the data types and relationships that the software vendor has
defined for you.
The ability to define relationships between data types in the system. Data Types are not standalone entities.
They will have relations to other data types such as master-detail type relations. An example of this is a “master
Sample” that may have many “Aliquots” under it. This type of relationship should be easy to configure in the
LIMS with no programming.
The ability to support one to many and also many to many relationships between data types. Support for
complex relationships is a nontrivial exercise for most applications in any domain. The LIMS should enable these
relations to be easily defined with no programming. These should also be changeable at any time by authorized
administrators.
The LIMS should provide the ability to easily define and arrange these data fields as well as group them together
on forms. Data fields describe Data Types and are what the user enters data into on forms. Examples include
“Sample Id”, “Sample receipt date”, etc. There should be support for a variety of data fields such as Date, Pick
lists, Text, integer, check boxes, etc. These should also be changeable at any time.
This list of features should all be done via a configuration. This means that no programming is involved and no Database
Adminstrator is needed. These types of changes should be doable by the lab manager with little or no training on the
LIMS. This ensures not only rapid project implementation timeframes and drastically reduced costs, but long term
viability of the LIMS to quickly adapt to new technologies that come into the lab.
To schedule a detailed technical demonstration of how Exemplar LIMS measures up to these requirements, please
contact us at [email protected] for more information.