70-U6900-26-CP343_081026 Leica SS.pdf.PDF

UNM Proposed Sole Source Form – Rev 04/07/2021

Proposed Sole Source Purchase Form Pursuant to N ew M exico Procurem ent law , the UN M Purchasing D epartm ent w ill post your com pleted form on the UN M

Sunshine Portal and the State of N ew M exico Procurem ent w ebsite for 30 days prior to purchase of the goods/services.

I. GENERAL INFORMATION. PLEASE PROVIDE THE FOLLOWING:

Date of Request 07/10/26 Requisition Number (If Applicable) Request Submitted by: Michael Paffett Title Research Facility

Director Department UNMCCC ALMSR Shared Resources Email mpaffett@salud.unm.edu Phone 2-1437

Proposed Vendor Leica Microsystems, Inc. Amount $119,000 Vendor Address 10 Parkway North, Suite 300

Deerfield, IL 60015

Term of Purchase

One time purchase

Buyer Team - See Commodity list at http://purchase.unm.edu/department-information/comm-code-listing.pdf

Provide a basic description of goods/services to be provided:

AIVIA Image Analysis Software (Leica Microsystems) is a software that combines artificial intelligence-guided image analysis and visualization solutions for scientific discovery. AIVIA is designed to improve accuracy, streamline workflows, and enhance data exploration.

Why is this purchase needed?

AIVIA will enhance and facilitate modernization the computational infrastructure in the Advanced Light Microscopy Shared Resource (ALMSR) that supports over 100 of HSC and Cancer Center users. This highly relied upon resource has several million dollars of state-of-the art microscope instrumentation that generates large volumes of content-rich imaging data. At the same time, microscopes produce more specialized and larger data that require sophisticated image analysis tools. Our current lack of state-of-the art image and data analysis tools is limiting what our users can do with their datasets.

AI Access for All - AIVIA makes advanced data analysis accessible for all biologists - with no computer science expertise required.

Radically Simplified Segmentation - AIVIA’s AI-powered analysis capabilities leverage a biologist’s expertise to generate robust and reproducible segmentation results.

Total Freedom on a Single Platform - AIVIA’s powerful and fast 2-5D visualization and analysis unlocks all the value of your data - within a single platform.

AIVIA Apex is a comprehensive microscopy image analysis solution for researchers who need a variety of image analysis applications. Apex also provides microscopists with the flexibility to apply their own deep learning models from third party or open-source repositories.

Ideal for large research groups or core imaging facilities, AIVIA Apex's Floating License Manager allows a site to run their AIVIA licenses on any machine on their local network.

II. BASIS FOR SOLE SOURCE PROCUREMENT. CHOOSE APPLICABLE BOX(ES) AND

PROVIDE ADDITIONAL INFORMATION, AS REQUESTED:

☒ Proprietary item, technology or service only available from the proposed vendor. (Check box and describe proprietary component)

Although AI is not unique to image analysis, there are some key features of AIVIA’s deep learning modules and contractual service/support elements that are described below that support the sole source.

☒ Compatibility requirement with existing item, technology or service. (Check box and describe compatibility requirement)

AIVIA is file type agnostic, which makes it compatible with all conventional image files type and all of the microscopy industry file types such as. lif (Leica); .czi (Zeiss); .vsi (Evident) and .nd2 (Nikon) files. It also works with our existing Imaging restoration licenses HDF5 (Huygens) that are vital for post-processing work flows as well as numerous others making this platform highly versatile.

☒ Renewal of support/maintenance/subscription of software, technology or other intellectual property. (Check box and describe)

Although there are number of AI image analysis platforms available, long term software maintenance and end-user support of AIVIA are unmatched. For example, no other vendor was able to provide 10 years of software maintenance and unlimited end-user support. This was a critical determining factor that allowed the ALMSR to leverage to avoid future price inflation while maintaining current software version and continuity of unlimited end-user support for the duration of the software maintenance and support agreement. The competing vendor for this project was only able to guarantee 5 years of purchasable software updates and the end-user support was limited on a hourly basis per license per institute. Software maintenance and end-user support are major considerations as these elements will enable our user base to fully utilize this technology without interruption in software iterations and all associated features as well as product support for our user base be fully supported by the AIVIA technical specialists.

☐ Other Basis for Sole Source: Please describe below:

III. SUPPLEMENTAL DETAILS. PLEASE PROVIDE ADDITIONAL INFORMATION AS

REQUESTED BELOW:

Describe in detail the unique capabilities of the proposed vendor’s goods/service and/or personnel performing the work and why this constitutes the only source. Focus on what is unique about the goods/service and why no other vendor could meet your needs.

Features Included in AIVIA Apex (CellBio and Neuro):

o Cell Analysis recipes (2D and 3D data sets) for the spatial analysis of complex and hierarchical biological entities and environments.

o Can generate counts on a per compartment basis (e.g. vesicle counts per cell, cell counts per brain region, etc.) as well as distances between/intensity measurement per different compartments/objects.

o Neuron Analysis recipes (for 3D data sets created by Electron Microscopy or Light / Fluorescence Microscopy) and the Neuron Composer – a dedicated toolset that semi-automatically detects and traces neurons in 3D fluorescence data sets and segments soma, dendrite branches, dendrite trees, and spines.

o Can generate counts and spatial location of every object detected, counts per sub-object (e.g. numbers of dendrites per neuron, branch angles, branch points, numbers of neurons, numbers of spines per neuron/dendrite).

o Ability to train and apply AIVIA’s machine learning-powered Pixel Classifier for image enhancement and segmentation for radically simplified image segmentation of 2D to 5D microscopy data sets.

o Deep Learning and Python Integration to apply any of AIVIA’s pre-trained deep learning models for image enhancement and/or segmentation.

o Wide range of image analysis recipes including nuclei count and tracking, cell count and tracking, particle count and tracking, cell proliferation assay, neurite outgrowth, wound healing (phase contrast), stem cell colony detection (phase contrast), cell tracking (phase contrast), object detection (meshes), object detection (spots), object tracking and lineage tracing.

o Cell Analysis - Cellpose and Multiplexed Cell Detection with Cellpose generalist deep-learning algorithm o Phenotyper for machine-learning based classification with 2 automatic clustering options: k-means and PhenoGraph-Leiden unsupervised automatic clustering methods.

o Spatial insights with Marker Cluster Dendrogram, Summary Plot, Binned Scatter Plot, and Dimensionality Reduction (UMAP, PacMAP, t-SNE) o Batch processing of recipes trained deep learning models, and trained pixel classifiers.

o Semi-automated analysis and editing tools such as the Contour Editor, which simplifies the task of reconstructing complex 3D shapes such as organelles, neurons and other cells, Object Classifier, a machine learning powered tool for classifying objects and identifying new phenotypes, and the Track and Lineage Editor.

o 3D objects generated using the Contour Editor can be used to delineate regions of interest where users can apply any of the recipes, trained Pixel Classifier, or deep learning models.

o Pixel/object colocalization analysis (includes computation of Manders’ and Pearson’s correlation coefficient).

o Open over 40 microscopy file types, including extended support for Leica files (FALCON and TauSense data).

o Render multi-channel 2 to 4D images of up to 3 TB (raw size) with millions of detected objects/cells/neurons/tracks.

o Highly interactive visualization tools including the Ortho Slicer and the Clipping Planes to selectively reveal regions, objects, and relationships of interest in raw and analyzed mesh data.

o Search and Action bar to access all training materials and to command AIVIA to take specific actions.

o Image and object text tagging system ideal for collaborative work and to visually keep track of different objects/regions of interest.

o The Video Animator to unleash creativity to generate high fidelity video animations of up to 4K and 60 fps.

o Spreadsheet and Charts (scatter, bar, lineage, and track) features to explore and export analyzed results for further statistical exploration in dedicated third party software.

o Virtual reality integration for interactively navigating and exploring both the raw voxel data and segmentation results using advanced VR headsets

Describe the due diligence made to locate other possible sources including communications with other universities, communications with similar providers, web searches, yellow page searches, review of advertisements and trade publications, etc.

We requested vendors in this technology space that support the very specific needs for core facility image analysis and would integrate smoothly with our existing Leica, Zeiss and Evident image acquisition infrastructure. The ALMSR Director and technical director are content experts in this space and drew from their collective knowledge; image analysis experience; existing relationships with other imaging/AI experts at other University departments and clear understanding of the microscopy/imaging industry’s offerings to source this product.

List the other vendors who were contacted. Please describe the specs/qualifications/criteria that the other vendors were unable to satisfy.

The UNMCCC Advanced Light Microscopy Shared Resource vetted the selection of vendor by hosting product separate product evaluation workshops where both Leica Microsystems (AIVIA-APEX) and Carl Zeiss (ARIVIS) were invited to compete. Another 3D image analysis offering (IMARIS) was considered, but an existing license already exists on the HSC campus and does not have the DL capability offered in the Leica AIVIA APEX with Dev Mode perpetual license. Therefore, IMARIS was not considered for this software initiative. Each AI software platform was put through a rigorous evaluation by our imaging community with real data sets to evaluate. Although each platform was able to perform most tasks in similar fashion, there were key technical and product support aspects that led to the AIVIA being selected. Our community selected the AIVIA solution based on ease of use; friendly, yet sophisticated user-interface; data-accuracy; positive feedback from user testing the Cell Biology and Neuroscience analysis ML/DL applications. Secondly, AIVIA APEX software has no subscription component to access the full complement of deep learning (DL) AI features either in-cloud or local. Leica’s solution for DL is a perpetual license and not a subscription service compared to the Zeiss offering which is cloud only. As a cost-recovery facility, we can’t afford a long-term subscription-based pay wall for DL access and need to own the actual license to comply with our Centers fiscal strategy. Leica Microsystems is also able to provide a total of 10 years of software updates with considerable cost savings whereas the Carl Zeiss is only able to offer 5 years. This is an extremely advantageous criteria as we will be fully supported with semi-annual software updates and customer support services for all HSC users for the 10 year duration.

KEY SPECS

Perpetual AIVIA APEX license for the software package:

o AI first, 2-to 5D image visualization and analysis software with all the functionalities of AIVIA Go and Elevate (both CellBio and Neuro versions).

o Floating License Manager: runs AIVIA Apex licenses (one at a time) on any computer within recommended specifications that has internet access.

9 years of maintenance (after 1 year of included 1st year maintenance) including:

o Unlimited upgrades to the latest version of AIVIA o Technical and application support using self-guided materials such as tutorial videos, how-to-articles, detailed explanations on the AIVIA wiki, and much more o Technical and application support with our image analysis experts via phone, email or remote sessions AIVIA DevMode add-on - Perpetual add-on license for training your own deep learning models using AIVIA’s compute engine:

o Users can choose to train a new deep learning model from scratch or perform transfer learning by leveraging one of the existing AIVIA deep learning models or by importing models that are outside of AIVIA. NOT A SUBSCRIPTION AND CAN BE PERFORMED LOCALLY