Lab Automation,
From the Ground Up
A guide to evaluating, implementing, and optimizing
automation solutions across your lab
PLANNING
for Successful Lab Automation
SELECTING
Scalable Technologies
APPLICATIONS
in Drug Discovery,
Organoids, and More
LAB AUTOMATION
RESOURCE GUIDE
2 Lab Manager Lab Automation Resource Guide
Table of Content
From Planning to Practice: A Strategic Guide to Lab
Automation ...................................................................3
Planning for Lab Automation ..........................................4
Why Add Automation to Your Lab ...................................... 5
Automation in Regulated Labs ............................................ 6
Evaluating Automated Liquid Handling Solutions..................... 9
Three Tips to Determine Which Features You Need
When Shopping for AI Software .......................................11
Implementing and Optimizing Automation.....................13
A Three-Step Approach to Successful Lab Automation.............14
Accommodating Automation Technologies in an Existing Lab ...18
Upgrading Stand-Alone Automated Liquid
Handling Systems to Workstations..................................... 20
Scaling Your Laboratory Automation: From Basics to Blue-Sky .. 22
Securing Success in Your Automated Lab............................ 24
Applications of Automation ......................................... 27
Assays and Automation in Robotic Workstations................... 28
Using Automation and Machine Learning in Drug
Discovery and Development ............................................ 30
Deep Learning as the Future of Organoid Work ................... 32
3 Lab Manager Lab Automation Resource Guide Introduction
This eBook covers how to assess your lab’s readiness for automation, select
technologies that align with your workflows, implement and scale systems
effectively, and explores real-world applications in areas like drug discovery and
organoid research.
From Planning to Practice:
A Strategic Guide to Lab
Automation
Tools, tactics, and insights for labs exploring automation
at any scale
Lab automation is no longer just the domain of high-throughput screening labs or industrial-scale operations. Today, automation technologies—from liquid handlers to robotic workstations—are accessible to labs of all sizes, offering a powerful way to boost efficiency, minimize
errors, and expand scientific capabilities, whether for a single process or an entire workflow.
Successfully implementing automation requires more than just choosing the right tools; it
takes careful planning, thought-out implementation strategies, and a strong understanding of
what these technologies can realistically achieve. Labs must evaluate their readiness, consider
regulatory and operational constraints, and select solutions that align with both current and
future needs. Once automation is in place, ongoing monitoring and optimization are essential
to ensure these systems deliver lasting value.
Chapter One
Planning for Lab
Automation
Automation is revolutionizing labs worldwide, promising greater efficiency, reproducibility, and scalability. But before diving into automation, lab managers must understand
the fundamentals—why it matters, how to integrate it into existing workflows, and what
factors influence its success. Without proper planning, labs risk investing in systems that
create new inefficiencies rather than solving existing ones.
This chapter provides a foundation for making informed automation decisions. Readers
will delve into the advantages of automation, ways to assess automation opportunities,
considerations for maintaining compliance in regulated environments, and how to identify the best technologies for their current and future needs. From selecting AI-enabled
software to evaluating liquid handling systems, this section focuses on the strategic planning required to adopt automation that supports immediate needs and long-term success.
WHY ADD
AUTOMATION
TO YOUR LAB
A guide for lab managers
considering automation
Lab automation is a big decision, whether it’s upgrading a benchtop
instrument or fully automating a workflow. If your lab is repeating the
same tasks over and over, and staff are overworked or making avoidable
errors, it may be time to consider automation.
KEY BENEFITS OF ADDING
AUTOMATION TO YOUR LAB
• Boosts efficiency and throughput
• Allows staff to focus on more
meaningful and stimulating work
• Reduces human error for more
consistent results
• Lowers the risk of repetitive strain
injuries
TIPS FOR SUCCESS
You’ve given it a lot of consideration
and decided that automating your lab
is the right choice, but how do you
ensure it’s done right? From staff and
leadership buy-in to choosing the right
equipment, there are a few main things
to consider before a lab automation
project gets underway.
Communicate early and often
• Involve staff from day one,
getting their input and keeping
them informed
Start small (if needed)
• Modular instruments let you scale
up over time
Choose the right workflows
• Focus on high-frequency,
time-consuming tasks
Set clear goals
• What do you hope to achieve—
improved efficiency, staff
retention, consistency?
Don’t forget the software
• Does the instrument integrate with
your software and existing IT
infrastructure?
• Is the software user-friendly?
QUESTIONS TO ASK BEFORE
AUTOMATING
• What are your most repetitive tasks?
• How often are these performed?
• Is human error common in
these tasks?
• How many steps are involved?
• How much space do you have for
equipment?
• How will you implement and train
staff on new systems?
• What is the cost-benefit ratio?
• What software or IT infrastructure
is required?
Even when done well, automation brings big changes, but with careful planning, clear goals,
and strong communication, it can be one of the best decisions you make as a lab manager.
6 Lab Manager Lab Automation Resource Guide
Automation in Regulated Labs
When deciding whether automation will benefit your lab operations,
there are several things to consider
By Gina Hagler
Most labs are subject to guidelines and requirements of one
sort or another, including ISO standards for non-regulated labs. Because many organizations issue regulations and
oversee compliance in regulated labs, different types of labs
have requirements that vary in the particulars, if not the
intent. Nonclinical laboratories are subject to Good Laboratory Practice under Part 58 of Title 21 of the Code of
Federal Regulations. The Centers for Medicare & Medicaid
Services regulate all non-research testing performed on
humans in the US through the Clinical Laboratory Improvement Amendments. Contract research organizations (CROs)
require strict adherence to regulations because they not only
take on the regulatory responsibilities of their sponsors but
are subject to comprehensive audits by their sponsors, the
FDA, and other regulatory bodies.
Automating repetitive or highly specialized work may be a
way to remain compliant while improving efficiency and even
the bottom line. Still, when deciding whether automation will
benefit your lab operations, there are several things to consider:
1. Which processes in your lab would
you automate?
In general, it makes sense to automate the most repetitive
and routine work. Because of that, pipetting systems make
7 Lab Manager Lab Automation Resource Guide
a good starting point. Electronic pipettes produce accurate
results with less reliance on individual technique, while
multichannel pipettes make it possible to reliably transfer
samples to plates in less time. The more high-tech version of
electronic pipettes is an automated liquid handling system
that can be configured as an automated pipette with a robotic
arm and a small footprint. For some systems, the arms are
available from a variety of manufacturers. With others, the
arm is part of the system, but the consumables can be purchased from any manufacturer.
Processes that are time-consuming and demand precision,
even if performed less often, are another good choice for
automation. DNA sequencing is one such process. The
available automation options cover the most basic to more
complex operations. For large laboratories, the use of such
automation may make even more sense. For complicated,
interrelated processes with many discrete steps, automation
using robots can be used to create an elaborate system. In
this system, materials pass from station to station for tasks
that are performed by the robotic arms as instructed through
the operating software. In such systems, when a task is completed, the resulting materials are conveyed along a track to
the next station. The entire operation can be done without
physical interaction once the process begins.
2. Is the equipment you want in
compliance?
It’s easy to get caught up in the possibilities of automation,
especially with so many choices promising so much in the
way of efficiency and accuracy. Identifying the automated
product or process and checking to see if the equipment
states that it is in compliance with your regulations is an
important reality check that will narrow your plan to one
that’s feasible. If you purchase the equipment from a party
other than the original vendor, or if there is a question of the
equipment being in compliance, performing a validity check
will be necessary. This check can be performed in advance of
purchase and to your specifications. If, ultimately, the equipment you’d like is not in compliance, alternatives may exist.
3. How will you adapt to the changes
that automation will create in the
current operating environment?
Looking ahead and preparing your team in advance for the
changes that come with automation will go a long way to a
smooth transition. Discuss the parts of existing processes
that will be impacted. Decide on the reports that will be
needed to keep any multi-step processes on track. If the person directing the process on an automated part of the system
will not be in the same location as the equipment, determine
the plan for interacting with that person. Taking the time to
get feedback about which processes to automate, where to
place the equipment, and potential pitfalls envisioned by the
people in your lab will help you to bring automation into the
lab as a positive development.
4. What costs and benefits are
associated with this automation?
The cost of automation is more than just the cost of the
equipment. There will likely be training, maintenance, and
materials specific to the equipment. It may be necessary to
change the lab’s current configuration, and you may need to
install new ductwork, piping, electric feeds, or other dedicated fixtures. Placement of the equipment in a location that
does not disrupt the workflow but instead enhances it may
also require some creative thinking and adjustments. The
goal is to reap the benefits of automation through greater
ease for maintaining compliance, greater efficiency in the
work performed, a reduced margin of error, and savings in
morale or dollars.
8 Lab Manager Lab Automation Resource Guide
5. Will the benefits outweigh the
costs over the life of the project or the
equipment?
Meeting the operating budget is essential, but the cost/benefit consideration is about more than the cost of the automated
equipment. How well will this new equipment interface with
your existing equipment? If there will be a workaround by
IT to “make the system work,” you’ll need to add that to the
cost. You will also need to consider the life of the technology
(automation) and that piece of equipment. Is this the newest
technology on the market? Will equipment from the generation before meet your needs? What’s on the horizon? Could
you purchase used equipment? Depending upon what you
learn, it may make it easier to decide if now is the time to act.
Once you are certain the equipment has a useful life and
realistic price point, it’s time to consider whether this expenditure will compromise your ability to meet necessary
outlays such as payroll and fund expenses for materials and
other projects. A machine may be perfect for the automation
you have in mind, and it may not cost “that much” relative
to other options. However, unless there is a tangible benefit
from that purchase during the life of the project it is intended for, there is no reason to make the expenditure. You could
see how this purchase would play against other purchases or
savings that result from the purchase, perhaps in a different
area, to determine if the purchase as part of a larger automation package results in a significant benefit.
Automation holds great promise for those required to remain in compliance with regulations. The judicious use of
automation can also free personnel to work on more intricate
aspects of a project, attracting and retaining talented people
in the process. It can improve results while reducing or eliminating some categories of error. Automation can also reduce
costs, freeing funds for additional equipment and training
that will help the lab remain in compliance.
9 Lab Manager Lab Automation Resource Guide
Evaluating Automated Liquid
Handling Solutions
Liquid handling systems should be adaptable to future needs, and their
maintenance should be transparent and straightforward
By Brandoch Cook, PhD
A liquid handling system is often the centerpiece of laboratory automation. This can be the case for drug discovery
and validation in big pharma companies, infectious disease
detection in clinical labs, fulfillment of chemical screening and next-generation sequencing (NGS) workflows in
university core facilities, and big-data-driven experiments
requiring many samples and replicates in individual research
laboratories. Although vastly different environments with
disparate concerns and goals, the through-line is that delegation of a subset of laboratory tasks to a robotic platform
imparts both efficiency and reproducibility not achievable by
human hands.
For instance, think about yourself or your laboratory personnel performing comparatively menial tasks such as nucleic
acid extraction and purification. Often, optimizing these
workflows dictates coordinating sample numbers with the
capacity of a laboratory centrifuge. As you have discovered
the hard way, the time to complete a round of extraction
builds exponentially rather than linearly when you have 24
10 Lab Manager Lab Automation Resource Guide
or 48 samples to wash, aspirate, and spin down repeatedly.
Additionally, the magnified possibility of a single labeling or
transfer error extending throughout the whole procedure can
be exceedingly difficult to track backward and identify.
Now, think about designing and executing large-scale experiments to generate complicated data sets using precious
reagents such as antibodies or recombinant proteins. An
unidentified manual error can have a catastrophic impact on
budget, time, and overall project success. Moreover, large
data sets intrinsically require extra technical and biological
replicates to reach increasingly stringent statistical significance thresholds. Automating liquid movement for repetitive
tasks provides fine control over volumes, ensures error-free
distribution across hundreds of samples, and minimizes
reagent waste. Most importantly, however, liquid handling
systems can miniaturize reactions beyond what we can physically accomplish, including reproducibly filling 1,536-well
plates in sub-microliter volumes.
This advance has been indispensable to the modern era
of biomedical science. Liquid handling systems can now
be scaled from industrial high-throughput chemical and
antibody screening or protein structural characterization
platforms down to benchtop decks capable of integrating
complete or partial automation into everyday laboratory
workflows—such as NGS library prep, qPCR setup, and
comparatively simple reformatting of liquid reservoirs into
microwell plates.
The first reflexive concerns when considering the purchase
of a liquid handler will be money and space. Given your
current and projected budget, what is the timeframe for a
return on investment for the instrument and its appendages?
Can you physically fit everything into your workflows for the
next several years while accounting for both the instrument’s
direct footprint and the negative space around it to ensure
freedom of movement and user safety?
The first key to being able to answer these questions is to
begin a relationship with a product vendor representative
who can develop a thorough understanding of your current
and future project needs. A vendor should be able to predict
the size and extent of instrumentation based on the numbers and types of samples you need to process, compatible
external devices you will need, and how the system should
be adaptable to evolve with future needs. For instance, if you
need to add an entirely new assay or workflow in six months,
how straightforward is that, and does your service plan cover
the new customization?
The vendor should be prepared with specifications that
extend beyond footprint and power consumption. These
can include whether a handler’s coefficient of variation is
sufficiently small across a range of volumes to meet your
workflow needs, which assays, kits, and reagents are compatible and/or validated with the instrument, how they ensure
management of quality and regulatory compliance, and
whether the void or dead volumes translate into quantifiable
long-term savings. Finally, the vendor should be able to provide a menu of service contracts tailored to your customized
setup and offer real-world estimates of expected downtimes
during maintenance.
“Automating liquid movement
for repetitive tasks provides fine
control over volumes, ensures
error-free distribution across
hundreds of samples, and
minimizes reagent waste.”
11 Lab Manager Lab Automation Resource Guide
Three Tips to Determine Which
Features You Need When
Shopping for AI Software
Defining your goals, future-proofing, and ensuring accuracy are essential to
finding AI solutions
By Gail Dutton
The right artificial intelligence (AI) enabled laboratory
software can enhance productivity and efficiency, but with
so many options, choosing the best solution isn’t necessarily straightforward. Keeping a few key points in mind can
detangle even complex and conflicting options.
AI is becoming a staple resource among scientists and offers
a range of capabilities. As a lab manager, it’s important for
you to understand what AI can and cannot do well, along
with what your lab needs today and in the next few years.
12 Lab Manager Lab Automation Resource Guide
Here are three tips to help you select the best software for
your intended uses:
1. Define your goals: Identify the goals of the software as
well as the automation for your lab. “This includes considering the tasks that need to be automated, the workflow,
and the expected outcomes,” says Neil Harper, founder
of PDH-Pro, a continuing education provider. Speech
recognition and conversational AI to streamline human-to-machine interactions, chatbots for data searches,
image recognition for pattern matching, machine learning
(which requires human training in your lab), and deep
learning using neural networks (which access data from
multiple sources and learn similarly to humans) are distinct
options. Understanding your goals and the capabilities of
various AI solutions will help you select the best applications for your uses.
2. Future-proof your investment: As your lab uses AI,
its potential applications will likely increase. Therefore,
choose an AI platform that can scale to handle additional AI-enabled data sources and workflows.
Also, ensure the AI software is compatible with your
existing and planned computing infrastructure. Depending
upon the organization, computer hardware refreshes
occur approximately every three to five years, according to the Uptime Institute. In the past year or so, many
universities and organizations have begun upgrading
their computing infrastructure to support AI applications.1,2
Involve your IT department in decisions to ensure the AI
software you are considering is compatible with the existing computing infrastructure in your facility, and discuss
any planned or likely IT upgrades to minimize the risk of
future incompatibility.
3. Consider accuracy and ease of use: Generative AI is
expanding into the life sciences with software that generates synthetic data in mere minutes, using only a few
lines of code and models trained on genomics, medical,
and other datasets. Generative AI is easy to use and
can augment otherwise limited data sets—just be sure its
results are accurate. When making your selection, weigh
ease of use and output accuracy against the time-consuming historical process of training machine learning
applications on your own data.
Selecting the right AI software is challenging but needn’t
be daunting. Keeping these three tips in mind can help you
identify the features you genuinely need today and help
future-proof your software to achieve—and maintain—optimal performance.
References
1. “Universities Make Upgrades to Connectivity via
Wave 2 Wifi.” https://edtechmagazine.com/higher/
article/2018/01/universities-make-upgrades-connectivity-wave-2-wi-fi
2. “Industry-University Partnerships to Create AI Universities.” https://www2.datainnovation.org/2022-ai-universities.pdf
Chapter Two
Implementing and
Optimizing Automation
Adopting automation is only the first step—ensuring it functions efficiently in the lab
requires thoughtful execution. Poorly integrated automation can create more problems
than it solves, disrupting workflows, creating bottlenecks, and limiting scalability. To
maximize impact, labs must align automation strategies with their operational goals,
select technologies that complement existing processes, and establish a framework for
continuous improvement. Implementation is not just about setting up equipment—it also
involves change management, staff engagement and training, and ongoing evaluation to
keep systems running smoothly and effectively.
This chapter will guide you through the process of successfully implementing and
optimizing automation in your lab. Topics include a structured approach to automation
implementation, strategies for integrating new technologies into existing workflows, and
insights into upgrading systems and scaling automation as lab needs evolve. With the
right planning and support, labs can move beyond initial adoption to create efficient,
flexible, and future-ready automated environments.
14 Lab Manager Lab Automation Resource Guide
A Three-Step Approach to
Successful Lab Automation
Navigate complex decisions and ensure a smooth transition with proper
evaluation, implementation, and communication
By Holden Galusha
Laboratory automation is often seen as a silver bullet for
reducing errors and cutting costs. But determining whether
you should invest in automation is a multivariable equation,
and the output is not a binary yes/no. The answer hinges on
numerous decisions.
The first step to determine if a lab should invest in automation is evaluating its needs, expected return on investment
(ROI), and safety considerations. This phase is where the
bulk of decision-making and analysis will happen.
Evaluation
Evaluating ROI
Having an acceptable ROI is vital to getting leadership on
board with introducing automation. It should be the cornerstone of your argument. While the specifics of what constitutes a good ROI will be different for every organization, the
general parameters dictate:
15 Lab Manager Lab Automation Resource Guide
• You can afford the costs of the equipment and infrastructure upgrades
• Automation won’t compromise quality
• The savings yielded from automation will pay for it in a
reasonable timeframe
By fulfilling at least these requirements, you have the foundation for an airtight argument.
Note that automation shifts the costs associated with lab
processes from operational expenditure (OpEx) to capital
expenditure (CapEx). Capital expenses are large, one-time
purchases with long-term horizons in mind, which demand
more money upfront and limit flexibility to adjust the system
after purchase. Processes carried out from OpEx have lower
financial barriers in the short term and are more flexible.
Lab leadership must take this shift into consideration. “With
purchasing and installing new automation equipment, we
always have the upfront cost associated with it, but over time
. . . the cost of automation comes out to be cheaper,” says
Meghav Verma, product manager at the National Institutes
of Health. “However, this is not a given, and we always have
to go through the evaluation process of understanding the
long-term gains. We also have to consider the time and
opportunity cost saved when employing automated solutions.
These costs add up in the long run.”
Maintenance and service fees should also be considered.
Such fees remain in OpEx because they are recurring,
but they should be accounted for in the evaluation phase
to ensure that the lab can justify all costs associated with
automation.
Evaluating laboratory needs
There is an inverse relationship between flexibility and automatability. If your lab’s processes are not standardized and
rely on in-the-moment judgment calls, then it’s possible that
your lab isn’t yet suited for automation. Automating prematurely can be more costly than sticking with manual processes until the processes have been better defined. Tesla Motors’
infamous Model 3 car production line is a prime example
of premature automation. Tesla designed an assembly line
with more than 1,000 robots to manufacture the vehicles,
but the robots couldn’t handle the changing geometry of
parts coming in different positions. Tesla had to hire several
hundred employees to keep up with their production quota,
undermining the point of all that automation.
Are processes in your lab too reliant on human intervention
to automate? If you’re not confident they can be automated,
work on standardizing them. Don’t neglect documenting
the processes either. Besides making onboarding smoother
for new staff, detailed documentation will be helpful when
you automate and translate manual processes into protocols
executed by machines.
Evaluating safety
While automation can positively impact safety by reducing
the risk of repetitive strain injury, in some cases, it can also
complicate it. As automation increases your lab’s throughput,
you may need to store higher quantities of chemicals in the
lab. It’s possible that this extra storage will push your lab’s
fire control zone past its maximum allowable quantity. Work
with the facility’s safety specialist to accommodate the new
equipment and chemicals.
In short, “a company is ready to introduce automation when
they understand the current process and its bottlenecks, they
can justify the investment, and they’re ready to support the
equipment,” says Hayden Allen-Galusha, automation engineer at Kumi Manufacturing of Alabama.
Implementation
After opting for automation, the implementation phase
begins. There are a few things to consider in this phase:
where to start automating, using project management tools,
and training.
To begin automating, look for the bottlenecks in your lab’s
processes. Such tasks will take disproportionately more time
to carry out than others and often will be monotonous. An
example would be pipetting. Rather than having an employee pipette dozens of times per day by hand, purchase
an automated pipettor. From there, you can start expanding
outward, streamlining more processes in your lab as you
learn how to identify opportunities to automate.
Software automation can also be a good starting point, as the
barrier to entry is much lower than hardware automation.
However, it still teaches you how to identify automation
opportunities. For instance, rather than manually exporting
results from an analyzer with a USB drive, explore options
16 Lab Manager Lab Automation Resource Guide
for connecting the analyzer directly to an electronic lab
notebook so that results are relayed instantly.
Using project management tools
When you begin implementing automation, you’ll need
more than just emails to effectively plan, communicate, and
delegate all the associated tasks. This is where dedicated
project management tools come into play. There are a wide
variety of tools and techniques available, ranging from things
as simple as Gantt charts to collaborative web apps. Before
the implementation officially begins, carve out some time to
demo a few project management programs with other stakeholders. Make a group decision on which software best suits
your needs and strive to update it consistently throughout
the implementation.
Training
Robust training is essential to successful automation. The
workers overseeing the equipment must become familiar
with not only usage but also maintenance and troubleshooting. There will be a lot to learn, especially if multiple systems are added simultaneously. “There has to be consistent
training programs associated with new equipment to allow
the users some time to learn and get used to the system,
rather than feeling like they’re being thrown in the deep
end,” Verma says. Depending on the backgrounds of your
staff, cross-training may be an effective strategy. Those who
have carried over experience using automated equipment
from previous positions will catch on faster, and they can
assist in training other staff members.
The most important thing is to foster an environment of
open communication. Staff must feel comfortable seeking
guidance whenever needed. Don’t become too ambitious
with the training timeline, as staff may get discouraged if
they aren’t meeting your expectations.
Resolution
Successful automation can be seen as a three-way partnership between hardware, software, and people. To sustain this
partnership, consistent effort must be put in all dimensions.
For hardware and software, it’s straightforward: keep up with
preventive maintenance, update the software, and, ideally,
integrate both with your lab’s existing informatics platform.
But to properly address the most challenging aspect—people—requires consistent communication and trust.
Cultivating trust and communication
“It’s a misconception,” Allen-Galusha says, “that automation’s sole purpose is to replace human operators. Many
times, automation is used to make a process more efficient,
increase quality, or make a process safer with the same
amount of manpower.” As the lab manager, you should communicate that introducing automation will not affect your
workers’ job security and ensure that there are opportunities
for staff to step into higher-value, more strategic roles. This
will imbue their jobs with more meaning and assuage their
fears of replacement, boosting morale. Additionally, you
should remain open to feedback and ideas. Your staff are
the ones most familiar with the newly automated processes, so they will have insight into how the processes can be
improved.
Staff must also trust the equipment. “The onus lies on the
lab manager to make sure the hardware and software is not
completely changing the way the scientists were performing their tasks,” Verma says. “The equipment should feel
familiar, which will increase adoption. Trust is an important
aspect of this, where users need to build that trust by verifying the data.”
Ultimately, automation is no silver bullet. Efficiency and
standardization are the foundation of a fruitful automated
system. In the words of Bill Gates, “The first rule of any
technology used in a business is that automation applied to
an efficient operation will magnify the efficiency. The second
is that automation applied to an inefficient operation will
magnify the inefficiency.”
Preparation and planning
Evaluate your lab’s workflows to identify where
automation can improve efficiency, reduce errors, and
relieve staff workload
Define clear goals for what you want to achieve
through automation (e.g., time savings, accuracy, data
consistency)
Build expertise and support
Select subject matter experts who will lead training,
troubleshoot issues, and ensure systems are used to
their full potential
Prioritize continuous training to ensure all staff remain
confident and competent
Support continuous improvement by regularly
collecting feedback
Engage and prepare staff
Involve key personnel early in the planning and
decision-making process, including managers,
technicians, assistants, and other users
Organize a planning event to encourage open
discussion and allow staff to voice concerns and
contribute ideas
Manage the transition
Set expectations early by communicating what
will change, when it will happen, and how it will
benefit the team
Monitor performance closely by tracking key metrics,
collecting feedback, and adjusting workflows as
needed
Mastering Automation Integration
Strategies to support successful planning, implementation, and long-term use
Successful automation integration in a lab extends far beyond installing the relevant equipment and instrumentation. It involves critical
evaluation of laboratory processes, staff roles, and interactions. The following checklist can help guide your lab through a successful
automation implementation.
18 Lab Manager Lab Automation Resource Guide
Accommodating Automation
Technologies in an Existing Lab
Implementing automation requires collaboration and thoughtful planning
By Marilee Lloyd
Introducing automation, whether via robotics, smart lab infrastructure, or another tool, can enhance research speed and precision. Automated elements streamline routine tasks like sample
handling and data analysis, freeing up researchers for more
complex work. When creating a new lab environment or updating an existing one, the project must be planned with these tools
in mind to be successful. However, the ability to smoothly and
effectively incorporate tools that have not even been invented
yet into that same lab five years from now is just as essential.
No one has a crystal ball, but planners and designers play a
crucial role in creating an adaptive and flexible lab environment that can change to harness these tools and the ones that
follow them. For lab managers, creating partnerships with
experienced lab planners and collaborating early will create
substantial advantages in the flow and useability of their
lab and their ability to introduce, evolve, and replace their
automated equipment and processes.
Collaboration first
Before diving into the specifics of any lab design, a laboratory planner must understand the nature of the research to be
conducted. This requires a deep and collaborative dialogue
19 Lab Manager Lab Automation Resource Guide
with the lab manager. Different scientific disciplines have
varying requirements; planners need to understand the
science you are planning to do today as well as what you may
do in the future. Understanding the lab’s long- and shortterm goals is critical for establishing the right strategies to
maximize the longevity and return on the investment you
are making in your lab.
Throughout the design process, the lab planner will be the
nexus for collaboration between the engineers, architects,
and automation specialists. By asking the right questions and
listening carefully, a lab planner can lead the entire team to
meaningfully investigate potential automation solutions and
realistically consider factors like scalability, future flexibility,
and compatibility with the existing infrastructure. As artificial
intelligence evolves, labs will likely deepen their investments
in robotics to perform repetitive and/or dangerous tasks.
Introducing and leveraging these tools to greatest effect requires the design to be flexible enough to accommodate them.
Ensuring that diverse perspectives are heard and the future
considered will lead to more robust and holistic lab designs.
Integrating the tools
Bringing in automation or automated systems (often via
robotics) in clinical testing lines or research assays requires
in-depth involvement with mechanical, electrical, and
plumbing (MEP) engineering team members. It does you
no good to have a robotic automated system in one location
if, in five or ten years, you need it somewhere else, and your
electrical infrastructure cannot support the move. Selecting
a lab planner with in-house engineering team members can
yield future benefits. It takes careful planning to determine
where robotic systems (and their supporting services) should
go, how to keep everyone safe, and how to make sure they
interface with other lab instruments. As such, other specialists such as environmental health and safety (EHS) experts
and equipment installation teams should be actively involved
with planning to ensure that the robots are placed in the
safest and most effective locations. Picture a well-choreographed dance of lab equipment: implementing robotics is
all about creating a smooth and efficient workflow, and such
choreography is only possible with facility engineers, safety
specialists, and equipment installation teams working collaboratively to define the moves.
Building automation opportunities
Emerging technologies and automation are also appearing
on the infrastructure side of the lab as well. Monitoring
particulates, air flow, lighting levels, water purity, humidity,
and other environmental conditions are all becoming more
prevalent. This creates a more reliable maintenance structure and alerts both lab staff and maintenance crews when
something is outside of control limits. Systems are regularly
or constantly monitored, reassuring owners of cGMP or
other sensitive environments that the space will be within
tolerance of the research or production requirements. In
addition, these automated and potentially AI-driven systems
can anticipate issues, schedule repairs, and re-balance system
controls. Lab planners are central to these discussions, often
serving as the linkage between MEP engineers, designers,
and end users.
Optimizing resources
An essential part of integrating new technology—be it
robotics, automated systems, or traditional equipment—is
incorporating sustainable design principles to ensure that
lab spaces are not only adaptable to evolving research needs
and technological advancements but also highly efficient.
Laboratories, being dynamic environments, must seamlessly
integrate automation without compromising their flexibility.
Robust infrastructure, modular and moveable furnishings,
and layouts that allow easy reconfiguration of equipment
and workflows all promote flexibility and sustainability by
minimizing resource waste and energy consumption.
This forward-thinking, sustainable approach guarantees that
the laboratory remains at the forefront of scientific research
while minimizing the need for costly renovations and reducing its ecological footprint.
No matter the type of automation being introduced or
upgraded, the result should always be that the laboratory environment is safe, effective, resource-efficient, and adaptive
to change.
“Before diving into the specifics of
any lab design, a laboratory planner
must understand the nature of the
research to be conducted.”
20 Lab Manager Lab Automation Resource Guide
Upgrading Stand-Alone
Automated Liquid Handling
Systems to Workstations
Automated liquid handlers have become indispensable by virtue
of freeing operators for other tasks while providing consistency
and reproducibility
By Angelo DePalma, PhD
Many labs today employ automated liquid handlers as standalone devices, manually moving plates to storage, incubators,
and readers. Integrating liquid handlers with other features or
functions—that is, turning them into workstations—is attractive, but the path to fuller automation can be costly and lengthy.
A liquid handling workstation combines two or more
functions—such as dispensing, aspiration, storage,
incubation, shaking, reading, or lidding/unlidding—into a
single system. These workstations can be semi-automated or
fully automated.
Workstations can also be defined by their purpose rather
than their components. For example, a workstation may be
designed specifically for PCR workflows or to run the same
panel of biochemical assays repeatedly.
21 Lab Manager Lab Automation Resource Guide
The appeal of a workstation is that you get an efficient, easyto-use platform. The trade-off, however, is that this simplicity can sometimes limit flexibility.
Upgrade considerations
Upgrading from a stand-alone dispensing system to a workstation involves adding one or more functionalities occurring
upstream or downstream of the liquid handler. Timing plays
a critical role in this decision. Manual plate movement can
be acceptable when processes are isolated and include long
gaps between steps. However, automating these transitions
becomes more logical if the dispensing step is quick and
followed by another short step, requiring the worker to keep
an eye on operations.
Labs in upgrade mode have probably identified process inefficiencies for liquid handlers they already own. Labs in total
acquisition mode, for example, those that take on new projects, have the luxury of choosing a ready-built workstation.
There can be a temptation to over-specify workstations to
prepare for every possible future need. However, defining
both current and future processes can be challenging, and
the resulting systems will be large, expensive, and complex.
A more strategic approach is to prioritize flexibility—selecting systems that meet today’s needs but can grow with
evolving business requirements. This approach is especially
attractive to many start-ups or small to mid-sized businesses
looking to minimize their initial investment.
Upgrading, moreover, has become simpler than it was even
five years ago. Software is more flexible, and vendors are
designing their components specifically for integration and
automation. This allows for purchasing exactly what is needed today, with a smooth path toward upgrading later.
Plusses and minuses
Both upgrade and “one and done” approaches have advantages and disadvantages.
Buying everything at once simplifies procurement and
ensures compatibility between components. It can also result
in smoother customer support down the line. On the other
hand, piecing systems together requires automation savvy
but affords less experienced labs the luxury of learning how
each piece works before investing in the next step.
Experienced labs with a good grasp of what is needed to automate an assay generally purchase everything needed from
the start. They also tend to anticipate better how their fluid
handling tasks will change over time and purchase integrated
systems or components accordingly. In contrast, less experienced labs often fail in their automation goals because they
perceive the problem to be less intricate than it actually is.
Labs should always become acquainted with an assay’s
manual operation before considering automation. Only after
breaking an assay down to its components can a manager decide the level and extent of automation required to improve
productivity. The worst strategy is attempting to automate
every operation through one purchase.
Flexibility is the key
Lab managers must consider every step in daily routines.
For example, does DNA extraction occur before PCR? Do
samples need to be prepared for mass spectrometry? There
are several ways to approach this, so workstations must offer
flexibility to accommodate them.
Before upgrading stand-alone liquid handlers to workstation
status, it’s important to identify frequently occurring, repetitive tasks that might be causing bottlenecks. Lab managers
should also ask themselves which tasks are growing and complex and justify automation. Not taking stock in this manner
often leads to failure.
Automation is not all about throughput and walkaway time.
Reproducibility and uniformity are necessary for tasks
where pipetting accuracy is critical. If five people work on
the same pipetting task, a robot and automated liquid handler will always be more reproducible and consistent.
Regardless of the acquisition path, deploying lab automation
intelligently and cost-effectively is difficult for resource-constrained laboratories. Managers may fret over potential
missteps with their first and subsequent automation components. For these labs, it’s best to err on the side of flexibility
and upgradability.
Once a workstation is installed, lab managers frequently
identify further optimization points to accommodate new
assays, which will require integrating additional devices or
adding modules to the liquid handler. A system that seemed
promising at first can quickly become limiting if it lacks the
adaptability to accommodate even minor changes.
22 Lab Manager Lab Automation Resource Guide
Scaling Your Laboratory
Automation: From Basics
to Blue-Sky
Questions to ask yourself when scaling up your laboratory automation
By Michael Schubert, PhD
Taking the first step into lab automation can be difficult—
but the challenges don’t stop when your first robot is online.
Similarly, even the most innovative automation platform is
no silver bullet if needs and processes aren’t clearly defined.
For laboratories interested in scaling up their automation,
it’s crucial to consider not just your needs but also how your
plans will integrate or adapt to existing equipment, workflows, and work volumes.
What are your lab automation needs?
Before making scaling decisions, consider your end-to-end
workflow. Identify any bottlenecks or pain points that are
causing throughput issues—or that may limit throughput
after increasing your use of automated technologies. Consult
with individuals across all aspects of your lab’s operations,
from sample handling to regulatory compliance, to factor
their knowledge into your decision-making. Finally, consider
23 Lab Manager Lab Automation Resource Guide
your lab’s current and future testing prospects. High demand
for a specific analysis or workflow may indicate a strong candidate for advanced laboratory automation, whereas steady
demand across domains may suggest a need to prioritize
automating multiple functions or those used across multiple
workflows, such as automated liquid-handling instruments
rather than maximizing the throughput or sophistication of a
single system.
Envisioning your lab’s future can also help you determine
the best approach to scaling now. For instance, labs that plan
to scale further in a stepwise manner may choose modular options that can be upgraded or extended according to
needs and budgets. Labs facing significant time or workforce
pressures may choose fully automated workflows with minimal training requirements, select vendors who can provide
off-the-shelf protocols and integrations alongside extensive
support services, or opt for more heavily artificial intelligence-supported solutions.
What is your lab automation setup?
Not all laboratories can accommodate all workflows. Your
lab’s computing power, flexibility, and capacity will determine your lab automation options—from LIMS integrations to data storage and encryption. Existing equipment,
software, and even the design of your lab space can further
dictate your choices and must be embedded into your plans
from their earliest stages.1
Poor planning can lead to inadequate integration, complex or failure-prone workarounds,
or hidden costs and inefficiencies that negate the benefits of
scaling your laboratory automation.
To make sure you’re scaling in a way that’s right for your lab,
start by mapping out your existing workflows and processes.2
Understand how samples move through the lab physically,
how data moves through your systems, and how workloads
are distributed between the steps of your existing and anticipated processes. This will not only provide insight into the
areas where laboratory automation can confer the greatest
benefit but also highlight adjustments that can be made to
existing processes ahead of scale-up.
What can you learn?
Lab automation is a journey, not a destination. Once you’ve
implemented your chosen solutions, it’s vital to continue
monitoring lab functions.3
Are your upgrades meeting all of
your lab’s needs and allowing you to achieve key measures
of performance? Has resolving one bottleneck introduced
another elsewhere? Are your new instruments interfering
with existing systems or other considerations like traffic flow
or ergonomics? Labs aiming to maximize the benefits of laboratory automation should regularly revisit their operations
and look for enhancement opportunities, especially if bugs or
issues arise. By engaging in continuous improvement, your
lab can gain insights into how you use your systems, where
challenges and pain points may arise, and how you can take
the next step into scaling your lab automation.
References
1. “The Future of Lab Automation: Opportunities, Challenges, and Sustainable Design Solutions.” https://www.
labdesignnews.com/content/the-future-of-lab-automation-opportunities-challenges-and-sustainable-design-solutions
2. “Planning for Laboratory Automation.” https://labcon.csmls.org/wp-content/uploads/2024/08/Planning-for-Laboratory-Automation-Presentation-Labcon-2024V5.pdf
3. “Embracing the Future of Lab Automation: Strategies for
Success.” https://www.linkedin.com/pulse/embracing-future-lab-automation-strategies-success
24 Lab Manager Lab Automation Resource Guide
Securing Success in Your
Automated Lab
Anticipating risks, failures, and scaling needs to realize
the full value of lab automation
By Michael Schubert, PhD
Laboratories are increasingly turning to automation to
increase quality and efficiency while minimizing risk and
error. Guidance abounds for labs taking their first steps into
the world of automated workflows—but for those already on
board, information is scarce. Where are issues most likely
to arise? How can risks be managed? And when is it time to
consider taking your automation to the next level?
Common failure points
Many of the most common error introduction points in lab
automation take place at the interface between humans and
computers. For example, one laboratory whose workflow required manual test requisition entry and communication of
results found that most failures occurred at those points.1
In
25 Lab Manager Lab Automation Resource Guide
fact, after automating their clinical biochemistry and immunoassay testing, the lab saw an unexpected spike in errors—a
fact they attributed to the need for a longer staff adjustment
period. Other laboratories have reported issues with insufficient staff training, poor troubleshooting guidance, and
accommodating new technologies in existing spaces not
designed for them.2,3
Within the systems, most errors occur in solenoids or sensors, which track specimens’ progress through the system.
When these sensitive detectors become dirty, faulty, or
misaligned, they may incorrectly report misplaced samples
or even halt the system until the error is manually resolved.
This may be as simple as realigning the part or as involved as
fully replacing it.
Errors are also common in the barcode-reading process.
Automated systems may struggle with poorly printed or
adhered barcodes, potentially requiring new suppliers or
printers for specimen labels. The readers may become dusty
or smeared, resulting in incorrect barcode readings, or they
can fail entirely due to laser misalignment. Even a tube that
isn’t fully vertical—whether because it was placed crookedly in its carrier or because of inconsistent spring elasticity
between carriers—can cause a barcode read failure.
Gripper failures, though not infrequent, are typically easy
to resolve. Most gripper-related errors arise from tube
misalignment in the carrier (causing the gripper to miss the
tube), labels separating from tubes and adhering to grippers,
or wear and tear on the pads. Fortunately, gripper motor
failure is much rarer.1
Rarest of all are instrument errors related to communication, although the results of such errors—specimen pileups
or system halts—have a disproportionate effect on overall
downtime. Most such errors are preventable (insufficient
water or reagent supply is a common culprit); some, however, are unavoidable. In general, devices with many moving
parts present a higher overall failure risk—but with careful
preventative maintenance, downtime is often minimal.
Staying safe
In terms of safety, lab automation is best known for its ability
to minimize people’s exposure to hazardous materials and
reduce the strain of repetitive tasks. Although automation
has risks of its own, pre-emptive action can help labs mitigate potential dangers and maintain a safe workspace.
• Conduct a risk assessment: Before implementing a
new system in your laboratory, understand the risks and
hazards it may introduce and determine if and how
these can be reduced or eliminated. Assign responsibilities as needed.
• Provide in-depth training: Ensure that all users understand how the system works and what to do in the
event of an error. Make training and documentation
easily accessible. If possible, aim to retain some familiarity in tasks and processes.
• Contain systems appropriately: Automated systems still
require manual interactions, which means that hazardous exposures are still a possibility. Employing suitable
enclosures or clean environments can reduce risk without impacting workflow.
• Avoid overcrowding: Automation may reduce the
need for people to move between lab stations but risks
increasing the number of people in small areas.2 Plan to
space analyzers so everyone has room to work without
wasting space.
• Provide protection: Large or complex equipment can
increase workplace exposure to noise and heat.2 Ensure
that staff have access to hearing protection (such as
earplugs or noise-isolating headphones) and that a
comfortable temperature is maintained.
26 Lab Manager Lab Automation Resource Guide
Outgrowing your automation
You’ve established contingency plans for equipment failure
and risk mitigation strategies to promote lab safety, and now
you’re wondering what’s next in your drive for continuous
improvement. How do you decide when to scale your laboratory’s automation, and how should you go about it? Many
of the signs to scale up will mirror those that motivated the
initial move to automation:
• Your test volume exceeds your staffing, time, resource,
or skill capacity
• You want to increase accuracy or reduce error (in existing manual processes)
• You want to increase speed or efficiency (for a specific
process or overall)
• You want to increase redundancy to reduce downtime
The needs driving your decision to expand can also help you
determine how to scale. Does it make more sense to increase
the throughput of already automated processes or to expand
your automation coverage to new processes? The former may
offer a lower entry barrier because you’ve already made a
successful case for automating those processes and can use
real-world data from your laboratory to support the need for
expansion. However, it can also present obstacles—the need
to find new tools that integrate with or easily replace existing
ones, the need to justify “upgrades” (and their associated costs)
soon after implementation, the need to retrain staff, and more.
Introducing new areas of automation carries its own set of
challenges. Priorities must be identified, and new purchases
must be justified. Systems must be accommodated in existing
lab spaces and integrated with other equipment. You’ll want to
pay careful attention to future-proofing: Which areas of your
laboratory are growing fastest? What tools and systems offer
interoperability? How can you build flexibility and redundancy into your expanding workflows? And, of course, each leap
forward in your laboratory’s technologies necessitates advocacy, training, problem-solving, and change management.
Leveling up your lab automation isn’t easy, but the advantages are worth the effort. With conscientious planning, solid
redundancy and repair strategies, and thorough risk management procedures, labs can leverage automation to realize
their full potential.
References
1. “Implementing a Laboratory Automation System: Experience of a Large Clinical Laboratory.” https://journals.
sagepub.com/doi/10.1177/2211068211430186
2. “Advantages and limitations of total laboratory automation:
a personal overview.” https://www.degruyter.com/document/doi/10.1515/cclm-2018-1323/html
3. “Challenges and Opportunities in Implementing Total
Laboratory Automation.” https://academic.oup.com/
clinchem/article-abstract/64/2/259/5608875
Chapter Three
Applications
of Automation
Automation does more than just speed up tasks; it’s changing how scientific work gets
done. These technologies enable unmatched levels of precision and deeper insights,
empowering scientists to tackle ambitious research questions with greater confidence and
reproducibility.
This chapter highlights real-world applications of automation, showing how today’s labs
are leveraging robotics, AI, and machine learning to advance research across diverse
fields, from drug discovery to organoid research. These use cases offer a clearer view of
where automation is already making an impact and provide a glimpse into how continued
innovation in automation will shape the future of scientific advancements.
28 Lab Manager Lab Automation Resource Guide
Assays and Automation in
Robotic Workstations
These technologies improve a lab’s output and save on priceless resources
By Mike May, PhD
The evolution of robotic workstations resembles that of computers. Gargantuan systems that only experts could operate
gave way to smaller, more user-friendly ones. Despite the decreasing size and simplified use, today’s robotic workstations
often outdo their predecessors, thanks to ongoing technological improvements.
A decade or so ago, automated liquid handling conjured up
images of room-size systems at pharmaceutical companies
costing hundreds of thousands of dollars and run by teams
of experts for operation and programming. Today, less than
$10,000, enough bench space for a microwave oven-size device, and some taps on a graphical user interface can get most
any scientist going in automated liquid handling. A huge
workstation handles far more samples, but that’s not needed
in most basic research labs. In fact, some scientists turn to a
do-it-yourself approach to automate processes in a lab.
Although life science and commercial labs primarily use
robotic workstations for liquid handling, that’s not the only
process that can be automated. These platforms can also heat
or cool samples, seal multi-well plates, and more. One team
of scientists turned esterase-based biosensors and a robotic
workstation into a pesticide-detection system, and the team
29 Lab Manager Lab Automation Resource Guide
reported that “a robotic system can be easily integrated in
industrial production lines, improving the monitoring efficiency, as well as the use of real-time biosensing devices for
environmental detection.”
When it comes to the basic reasons to automate a workstation,
most scientists know that this technology can improve a lab’s
efficiency. Plus, reducing human intervention leads to fewer
errors and variability in experiments. Despite those benefits,
some labs get more out of this technology than others.
Robotic workstations excel in situations with unchanging
workflows, such as clinical, forensic, and analytical service
labs that run the same tests or assays. These labs also benefit
from tracking samples and how they were treated, which are
two of the strong points of a robotic workstation.
Nonetheless, automated workstations’ capabilities keep
growing. As access to this technology expands to more labs,
the applications and modifications will expand as well.
Enhancing the advancement
More than the parts of a workstation matter when it comes
to what it can do. In some cases, advances in one area spawn
improvements in another. Automation and assay technologies are in a kind of perpetual feedback loop—as one group
progresses, the other responds with its own advancements.
The results of those advancing steps let scientists explore
more complex questions, often in more precise ways. For
instance, advances in automated liquid handling technology have focused on improving accuracy at a wide range of
volumes—from nanoliters to microliters. The ongoing trend
of miniaturizing assays to use less sample requires the ability
to work accurately with very small volumes.
The control of workstations also keeps improving, making
systems more approachable and preventing inadvertent
process changes.
Exploring the economics
Expense comes to mind when any lab manager thinks about
an automated workstation. In the days of gigantic systems,
the cost of robotic liquid handlers far surpassed the budgets
of most labs. Today, some scientists think that automated
systems include an economic incentive, but that’s not necessarily the case.
The economic benefits of robotic workstations are often
misunderstood. An automated platform will probably require
the same amount of consumables and reagents—maybe even
more—than manual methods. Automation has economic
benefits over its lifetime, which come from reduced retesting,
faster sample accessioning, and improved data integration.
An automated workstation, though, can also save labs money
in other ways. Besides capital equipment, one of the costliest resources in a lab is its personnel. It is not cost-effective
to increase personnel in response to increases in sample
processing. To handle more samples, a robotic system—an
affordable one—could be a lab’s better choice. Such a system
could even save labs money in less obvious ways by reducing
or eliminating the risk of repetitive strain injuries that can
slow individual or lab-wide progress.
30 Lab Manager Lab Automation Resource Guide
Using Automation and Machine
Learning in Drug Discovery and
Development
Automation and machine learning are transforming traditional workflows in
drug discovery
By Morgana Moretti, PhD
In the drug discovery race, pharmaceutical companies face
a common problem: traditional candidate screening is slow,
costly, and labor-intensive. To address these challenges,
many have adopted automation and machine learning (ML).
Together, these technologies optimize efficiency, reduce
costs, and increase drug discovery success rates.
Applications of automation and AI in
drug development
The core steps in developing a new drug have remained
consistent. First, a target within the body is selected for the
drug to interact with. Next, researchers design a molecule to
affect that target. Once synthesized in the lab, the molecule
31 Lab Manager Lab Automation Resource Guide
undergoes lab testing to confirm it works as intended and
does not produce unwanted effects. Finally, it is tested in
humans to assess its safety and effectiveness.
Automation and artificial intelligence (AI) can optimize each
of these steps. However, most companies are focusing on
three failure points in the drug development pipeline:
1. Target identification and validation
2. Drug design
3. Patient stratification for treatment benefit
Target identification and validation
ML enables the analysis of vast genomic, proteomic, and
transcriptomic datasets. For example, it can analyze patient
data to identify genes and proteins that are consistently
overexpressed or mutated in certain conditions, providing
new leads for drug development. Once a target is identified,
ML assists in validating its role in disease, determining how
a drug can modulate it, and predicting potential off-target effects.
Koon Mook Kang and colleagues recently described an
application of ML in drug discovery.1
They developed a
deep-learning model that predicts where drugs can bind to
proteins. Using this method, they identified a new binding
site on the P2X3 receptor, which allowed them to screen for
potential drug candidates. The model improved hit rates
tenfold and significantly accelerated the discovery of novel
P2X3-targeting compounds, which are important for treating
chronic pain and respiratory diseases.
Drug design
Designing a drug that effectively interacts with the target is
a primary focus of today’s innovation. ML can generate molecules with 3D structures tailored to interact with biological
targets. Moreover, scientists can use ML to redesign existing
drugs, enhancing their binding to disease-related proteins
or repurposing them for new therapeutic applications. After
making adjustments in simulation, researchers can synthesize and test the most promising designs.
A notable example of AI and automation in drug design is
the work of King-Smith and colleagues at the University
of Cambridge.2
They developed a platform that combines
automated experiments with ML to predict how chemicals
will interact. Their approach, validated on a dataset of over
39,000 pharmaceutically relevant reactions, is called the
chemical ‘reactome’. This approach can significantly accelerate the drug design process because it reduces the need for
time-consuming trial-and-error approaches in the lab.
Patient stratification
ML can uncover patterns that predict treatment outcomes
by analyzing large datasets, including clinical trial results,
genetic information, and patient health records. For example,
in a study of 2,538 cancer patients treated with atezolizumab,
an ML model used clinical and biological factors to group
patients into high- and low-risk categories for mortality,
providing more precise predictions for treatment success.3
This type of ML-driven stratification is valuable in optimizing treatment decisions and improving overall outcomes in
oncology treatments.
From challenges to change
Automation and ML have accelerated drug discovery by reducing costs and improving processes. However, limitations
remain, such as algorithmic biases, ethical concerns, and the
need for high computational demands. As these technologies
advance, it will be interesting to see how they will redefine
the operational, strategic, and competitive landscapes of
drug discovery.
References:
1. “AI-based prediction of new binding site and virtual
screening for the discovery of novel P2X3 receptor antagonists.” https://pubmed.ncbi.nlm.nih.gov/35849939/
2. “Probing the chemical ‘reactome’ with high-throughput experimentation data.” https://pubmed.ncbi.nlm.
nih.gov/38168924/
3. “Using machine learning for mortality prediction and
risk stratification in atezolizumab-treated cancer patients:
Integrative analysis of eight clinical trials.” https://pubmed.
ncbi.nlm.nih.gov/35871390/
32 Lab Manager Lab Automation Resource Guide
Deep Learning as the Future
of Organoid Work
AI-powered organoid analysis has proven effective, but what do next steps
look like?
By Holden Galusha
Organoids, a product of 3D biology, are cells grown to
develop a similar structure to their source tissues, in turn
mimicking the behavior and properties of that cell. Researchers often use fluorescence microscopy to observe the
development and morphology of organoids. However, this
manual analysis is time-consuming and leaves a high margin
for error. To address these issues, some researchers have
leveraged machine learning (ML), a form of artificial intelligence (AI) that humans can “train” to recognize patterns and
make more accurate predictions.
For example, a September 2021 study published in Development showcases MOrgAna (ML-based Organoid Analysis
software), an ML model designed for analyzing the morphological and fluorescence characteristics of organoid
models via image processing. Described as an “easy-to-use
ML pipeline,” MOrgAna represents an accessible path
to augmenting organoid analysis with ML. Rather than
analyzing organoids manually, researchers can instead
invest that time into training an ML model to perform the
analyses autonomously.
33 Lab Manager Lab Automation Resource Guide
Similarly, an October 2020 study published in Nature Communications details the use of ML to analyze 3D preclinical
colorectal and bladder organoid models and identify robust
drug biomarkers. The identified biomarkers accurately predicted patient responses to drugs used to treat colorectal and
bladder cancer.
These studies and others illustrate that ML was the next step
in organoid-centric research, and they have aided in developing more effective therapeutic treatments. However, conventional ML is not without its limitations. For instance, the
accuracy of ML models improves when fed new data, which
means that humans must intervene continually if they wish
to improve its accuracy. Additionally, ML models can be
time-consuming and costly to generate. A multi-disciplinary
expert is needed to design the model’s “feature extractor,”
which enables the ML model to identify points of interest in
data after being trained. In the case of 3D organoid analysis,
the model’s feature extractor would serve to identify cell
components, disease markers, and other characteristics in the
3D image. Implementing an ML model’s feature extraction
ability can be long and difficult.
To address these hurdles, some researchers have begun
using a variation of ML called deep learning (DL) to analyze
organoids—and it’s possible that DL will mark a new era in
organoid research.
Deep learning as the way forward
DL is a subset of ML that uses artificial neural networks—
that is, digital counterparts to the human brain—to learn.
Much like how biological brains are comprised of neurons,
artificial neural networks are comprised of “nodes” that are
stacked into “layers.” Simply speaking, the more layers a DL
model has, the more advanced it is. The main advantage that
DL models have over traditional ML models is that their
artificial neural networks allow them to effectively learn the
same way the human brain does. Consequently, DL models
can self-improve, are highly flexible, and are capable of identifying complex patterns and correlations that ML models
cannot perceive. In the context of organoid analysis, these
traits could manifest as being able to identify a wider range
of organoid types, identifying trends between organoid images more accurately, and effectively handling more complex
datasets than ML models can. For instance, a DL-based
organoid analysis solution called OrganoID was trained on
images of pancreatic cancer organoids but can accurately
identify organoid models of lung, colon, and adenoid cystic
carcinoma organoids as well. Traditional ML models likely
would not be able to replicate those results because they
require researchers to manually design unique feature
extractors for different organoid types rather than generating
their own feature extractors and then applying that extractor
to novel data. Moreover, researchers have demonstrated
that DL models can outperform conventional ML models in
image processing, which is the method by which organoid
analyses are carried out.
While DL models require a much larger set of initial
training data than traditional ML models, they can realize feature extraction autonomously—eliminating the
time-consuming and costly process of designing a feature
extractor for conventional ML models while yielding higher
accuracy. Though traditional ML models are still powerful
and well-suited to certain types of tasks, they cannot process
the same volume or complexity of data as DL models. This
naturally limits their analytic capability.
For these reasons, DL models open new doors in organoid
analysis, posits a paper recently published in Bio-Design and
Manufacturing.
The paper, “Organoids revealed: morphological analysis of
the profound next-generation in-vitro model with artificial
intelligence,” was written by researchers from the State Key
Laboratory of Bioelectronics at Southeast University, China,
and published in January 2023. DL, the authors argue, is a
fast-approaching future, with the body of literature around
the use of DL in biomedical contexts already mature. On its
current trajectory, the authors “anticipate more combinations of DL algorithms and high-content analysis.” Namely,
they “expect the development of a ‘self-learning microscope’
that would enable automated, high-speed, high-throughput,
highly repeatable analyses that could make organoid analysis
more comprehensive and easier to execute. “Luckily,” the pa-
“The main advantage that DL models
have over traditional ML models is
that their artificial neural networks
allow them to effectively learn the
same way the human brain does.”
34 Lab Manager Lab Automation Resource Guide
per states, “more and more biological and medical researchers are now devoted to organoid exploration and study. We
believe that readers will observe a significant increase in the
scope of DL-based organoid analysis in the future.”
While the future looks promising, DL analysis still presents its own set of challenges. For instance, most scientific
literature involving DL uses supervised learning, which requires manually annotated data (as opposed to unsupervised
learning, which does not require manually annotated data
but whose results are difficult to verify): “To achieve good
accuracy, neural networks usually require many annotated
samples to perform training tasks. Collecting annotated
high-quality datasets . . . for supervised learning is usually a
very difficult task, and manual annotation on these images is
also very tedious and expensive,” the paper says. While DL
models perform feature extraction autonomously, which is
a significant advantage over traditional ML models because
human intervention is not continually required, researchers
must still invest significant time into annotating the training
data. The authors propose the construction of standardized
databases for organoid images that any lab could access to
train their own models. Furthermore, DL demands very
powerful computers, and most current hardware cannot keep
up with those demands. As a consequence, the 3D organoid
images are effectively downscaled, which allows them to be
processed but can cut elements of interest from the image.
There is still work to be done, but the future is clear: the
trend of leveraging DL to facilitate organoid analysis is gaining traction, and we are likely to see substantial growth in
research around the topic. Ideally, this technology will make
organoids more effective and consistent, allowing therapeutic
solutions to be developed more efficiently than ever.
