Bay Area Chapter
Applications of AI/ML in the Automated Life Science Laboratory
LRIG Bay Area Fall Event - October 4, 2023
Embassy Suites, South San Francisco
Our next event will showcase the latest developments in AI/ML, data pipelines and more data science topics. Attendees will learn from leading experts on the role of data pipelines in drug discovery research, how labs use data science and new applications and developments in AI/ML.
Call for Posters: We are partnering with SLAS to sponsor an award for best poster by an early career researcher. The prize includes a SLAS2024 full conference registration and travel stipend. Topics for the poster do not need to reflect the AI/ML topic. Visit the Poster Session page for details. Please contact Maureen Beresini to submit an abstract.
The exhibition is now full. Vendor details are here.
As usual, the event is free, but please register so we can gauge attendance for the technical program and buffet lunch.
Agenda
11:00 Exhibits open
12:00 Lunch service
12:30 Welcome to LRIG, Louis Murray, President LRIG Bay Area
12:35 Introduction, Maureen Beresini and Dave Wexler, Program Chairs
12:45 Russ Berman, CTO at Eikon Therapeutics. Combining Advanced Engineering with Machine Learning to Advance Drug Discovery
1:10 Nicholas Larus-Stone, CEO at Sphinx Bio. Using LLMs to extract plate data from messy spreadsheets
1:35 Ryan Renslow, Director Computational Chemistry at Totus Medicines. Automated AI-based processing pipeline for DNA-encoded library screening.
2:00 Break
2:30 Will Van Trump, Sr Director Lab Operations at Spring Discovery. Using Machine Learning to Discover Inflammasome Therapeutics via High Throughput Phenotypic Screening
2:55 Michelle Li, CEO at Medra. Leveraging AI for Fully Autonomous Labs
3:20 Eric Wu, PhD student/ML Scientist, Stanford University and Enable Medicine. 7-UP: Generating in silico CODEX from a small set of immunofluorescence markers
4:00- End of technical program and event
Abstracts:
Russ Berman, Chief Technology Officer, Eikon Therapeutics
Combining Advanced Engineering with Machine Learning to Advance Drug Discovery
By combining Nobel Prize-winning microscopy, advanced engineering, automation, & machine learning, Eikon Therapeutics is pioneering a new way to discover drugs and understand cell function. We have built a fully automated platform capable of studying the motion of proteins in living cells, at a grand scale. By quantifying the movement of each individual protein molecule over time, we can build a profile of that protein’s population dynamics, both in its native state and in the presence of drug candidates, hormones, or other stimuli. We can capture hundreds of thousands of protein trajectories in less than a second, and we typically analyze millions of experimental conditions each week. Machine learning is used for everything from understanding the health of the cells to inferring whether the proteins’ function changes under compound treatment to predicting how proteins under study might interact with other proteins. This talk will provide an introduction to Eikon’s platform and its applications, with a high-level overview of how we use machine learning to help us better understand the biology of our targets.
Nicholas Larus-Stone, CEO of Sphinx Bio
Using LLMs to extract plate data from messy spreadsheets
LLMs have recently emerged as a versatile approach for a range of machine learning tasks, but have yet to be deployed at scale in the biotech industry. In this talk, we will show how to use LLMs to address a pressing issue in biotech data management: the extraction of plate data from unstructured, often messy spreadsheets. Both platemaps and results from plate-based assays are commonly stored in inconsistent formats, complicating downstream analyses and decision-making. We use LLMs to identify the location of the plate data even when these are embedded in poorly structured formats or accompanied by irrelevant information. Importantly, our method can handle a wide variety of plate layouts without requiring retraining, thus reducing the need for writing and maintaining bespoke scripts.
Ryan Renslow, Director Computational Chemistry at Totus Medicines.
Automated AI-based processing pipeline for DNA-encoded library screening.
DNA-encoded libraries (DELs) offer unprecedented screening capacities, ranging from
hundreds of millions to billions of molecules. At Totus Medicines, we have developed a
unique high-throughput cell-based DEL assay capable of screening multiple libraries
against an array of disease targets on one 96-well plate. This generates a deluge of
data for each screening run that consists of hundreds of billions of sequences.
To efficiently manage this data and expedite hit identification, we have developed a
robust, automated data processing pipeline. Deployed on Amazon AWS cloud
infrastructure, the pipeline is engineered using SQL and Python and performs real-time
data synchronization. It categorizes incoming data into long-term databases and
subjects the data to a comprehensive AI analysis. This automation accelerates the drug
discovery timeline and enables the rapid transition of hundreds of promising candidates
to resynthesis and validation within days of screening.
Will Van Trump, Sr. Director of Lab Operations at Spring Discovery
Using Machine Learning to Discover Inflammasome Therapeutics via High Throughput Phenotypic Screening
Many biological processes involve complex and convergent pathways, making target identification and drug action difficult to resolve with traditional approaches. In this presentation, we demonstrate the power of machine learning and advanced multi-dimensional data analysis in enhancing drug discovery. These methods enable the efficient analysis of complex, physiologically relevant data at high throughput and minimal cost.
With a single workflow, we were able to screen over 12,000 compounds to advance 300+ drug candidates over seven distinct targets. This was made possible by combining single-cell level imaging feature extraction and well-level highly multiplex proteomic evaluation. Every image was analyzed using a mixture of automated quantification of targeted cell biological features, unbiased discovery of functional cell subpopulations, compound similarity metrics, and deep learning signatures specific for various validated inflammasome cellular states. To complement our imaging data, we also performed multidimensional cytokine analyses on supernatants from the same inflammasome-activated plates that were imaged. Multiple scoring rubrics were designed to categorize compounds into novel inflammasome inhibitor classes whose mechanisms of action can be uniquely matched to relevant human pathologies.
Michelle A. Lee, Founder and CEO at Medra
Leveraging AI for Fully Autonomous Labs
As AI continues to revolutionize drug discovery, most lab automation still depends on pre-programmed protocols and schedulers. What if we could embed AI and machine learning directly within the wet lab itself? Medra is pioneering fully autonomous labs powered by computer vision and AI, offering increased flexibility, robustness, and generalizability for lab automation. In this talk, we’ll delve into how AI-driven systems can empower scientists to confidently scale their throughput, even for complex protocols such as cell culture and passaging. We’ll also look at how robotics and AI used in other fields, such as self-driving cars and manufacturing, can be used in the wet lab.
Eric Wu, PhD student/ML Scientist, Stanford University and Enable Medicine
7-UP: Generating in silico CODEX from a small set of immunofluorescence markers
Multiplex immunofluorescence (mIF) assays multiple protein biomarkers on a single tissue section. Recently, high-plex CODEX (codetection by indexing) systems enable simultaneous imaging of 40+ protein biomarkers, unlocking more detailed molecular phenotyping, leading to richer insights into cellular interactions and disease. However, high-plex data can be slower and more costly to collect, limiting its applications, especially in clinical settings. We propose a machine learning framework, 7-UP, that can computationally generate in silico 40-plex CODEX at single-cell resolution from a standard 7-plex mIF panel by leveraging cellular morphology. We demonstrate the usefulness of the imputed biomarkers in accurately classifying cell types and predicting patient survival outcomes. Furthermore, 7-UP’s imputations generalize well across samples from different clinical sites and cancer types. 7-UP opens the possibility of in silico CODEX, making insights from high-plex mIF more widely available.
Vendor Exhibition
Over 20 vendors from the life science community will be exhibiting at the event. Their sponsorship makes our program possible, so please be sure to visit. We extend a special 'thank you' to our premium sponsor Hamilton Storage.
Confirmed vendors as of September 20:
Hamilton Storage, Titian, FORMULATRIX, Dynamic Devices, BioNex Solutions Inc., Genedata, Thermo Fisher, Tecan, Tekmatic, Analytic Jena ,Opentrons ,GBO ,Complete Genomics ,BMG Labtech ,Automation Trainer ,Thompson Scientific ,Liconic Thomas Scientific ,IMCS, Inc. ,Araceli Bio
Chapter Officers
Louis Murray, President
louis.murray@lrig.org
Louis Murray has spent over 20 years in the Bay Area focused on Lab Automation across various roles, including; current North Americas Sales Manager with the Lab Automation group at Thermo Fisher Scientific responsible for all commercial activities. Automation Engineer at Exelixis, beta tested the GeneMachines RevPrep, Field Application Manager at Velocity11 building and supporting robotic platforms around the globe and Product Manager at Agilent Technologies for the automation portfolio and launched the Encore MultiSpan Liquid Handler. Louis joined the LRIG Bay Area chapter in 2021 and accepted the role of Executive Chair in 2023. Louis holds a Master of Science from the University of Aberdeen.
Michael Biros, Treasurer
Mike has over 30 years marketing and sales experience working for companies developing emerging technologies that accelerate life science research. He was the first marketing hire at LJL BioSystems, which was acquired by Molecular Devices in 2000. At Molecular Devices, he led the Drug Discovery instrumentation-marketing group as Director of Marketing, introducing several major fluorescent imaging- and point-reading systems. He was a principal and co-founder of Kallidus Group, a marketing agency serving life science, diagnostic and technology companies. He served as Vice President, Product Marketing at IntelliCyt and founded Cellerynt Group, a marketing agency focused on improving marketing automation for life science companies. He served as Executive Chair of LRIG, Bay Area Chapter from 2002 to 2023 before assuming the role of Treasurer. Mike holds an undergraduate degree in Biochemistry and Cellular Biology from U.C. San Diego, and an MBA from the Leavey School of Business at Santa Clara University.
Maureen Beresini, Program Co-Chair and Scientific Advisor
Maureen Beresini has over 20 years of experience in small molecule drug discovery at Genentech. She leads a group that develops and provides biochemical and cellular assays to promote the discovery of small molecule therapeutics. In recent years, Maureen has focused on high-throughput screening. In addition to the assays, she has contributed to the establishment of high-throughput automated systems and processes for conducting assays, analyzing data, and managing compounds. Maureen began her industry career with Syva Company, developing assays for therapeutic monitoring of immunosuppressive drugs. She received her doctorate in Biological Chemistry from the University of California Davis and was a postdoctoral fellow in the Cancer Research Institute at University of California San Francisco.
Dave Wexler, Program Co-Chair
Dave Wexler is currently Senior Director, Process Development & Automation at Exact Sciences based in the San Francisco Bay Area where he is responsible for leading the Assay Development, Process Development, Systems Engineering and Process Validation teams to design, develop, implement, test and validate processes and instrumentation critical to enable complex molecular diagnostic testing in a clinical Reference Laboratory. He provides technology vision and strategy to enable histopathology and molecular laboratory operations to scale and increase efficiency Prior to Exact Sciences, over the course of 13 years, Dave served in leadership positions with increasing responsibility at CareDx. The most recent being Vice President, Automation Engineering where Dave led a multidisciplinary group of Automation Engineers, Software Engineers, RA/QA and laboratory operations personnel to implement laboratory technology to increase throughput 10x for a Next Generation Sequencing (NGS) based diagnostic test.
Key accomplishment at CareDx: Dave directed the successful implementation and validation of Next Generation Sequencing (NGS) equipment, LIMS and data analysis software required to launch AlloSure®, a novel diagnostic test using cell-free DNA as a biomarker in a CLIA Reference laboratory. Additionally, Dave held a previous role as Senior Director, Clinical Informatics and Automation Engineering for CareDx. In this position, and earlier positions at CareDx and AGY Therapeutics, Dave was instrumental in the design and validation of new laboratory space for clinical and high-throughput drug discovery operations. Key to the ongoing success of CareDx, Dave worked closely with the head of Laboratory Operations to research, develop and validate new processes for increasing throughput and efficiency for processing of patient samples.
He received his Bachelor of Science degree in Chemistry from University of Denver, his Doctorate in Physical Chemistry from the University of California, Los Angeles and performed post-doctoral research at University of California, Berkeley.
Amer El-Hage, Secretary
Amer is an independent consultant in product development and project management for the transfer from development of life-science and laboratory instruments to quality manufacturing.
From 1991 to 2000 he was the Director of Engineering Programs at LJL BioSystems, and following its acquisition at Molecular Devices until 2004. Prior to LJL, Amer held engineering, manufacturing and project management positions at Beckman Instruments and Varian Associates.
Amer technical expertise is in instrument design and automation, micro-fluidics, plastic molding, Opto-mechanical, consumables and BioMEMS applications. He is versed in product design, verification and validation processes, ISO quality Design Control systems, and FDA regulatory compliance and CE/CSA/UL agency certifications. He currently chairs the Society of Bimolecular Sciences (SBS) Microplate Standards Development.
Mr. El-Hage’s name as an inventor is on fourteen US and two international patents. Amer co-authored journal articles and made presentations at international Laboratory Automation conferences and at academic institutions. Amer received his M.Sc. in Engineering with distinction from UC Berkeley in 1980.
Lisa Simmons, Board Member
Lisa Simmons is the Western Regional Sales Manager for Hamilton Storage Technologies where she is the sales product specialist for automated storage and retrieval systems. Lisa has over 20 years of life sciences industry experience in the area of benchtop automation and storage systems. for sample management. For 15 years Lisa was the key sales and marketing director for BioMicroLab (since acquired by STP Labtech) with robotic automation equipment for drug discovery laboratories. Lisa has held many roles from sales and marketing, account management, distributor channel management and production liaison. Lisa joined the LRIG, Bay Area Chapter in January of 2023. Lisa holds a Bachelor of Arts undergraduate degree from the University of Georgia and an MBA in Entrepreneurial Studies from Concordia University Irvine.