Imaging science, end to end. And the discovery biology around it.
Where is the gap? That is the step I take, not only the analysis at the end: methodology, instrument selection, fluorophore choice, immunopanels, live-imaging settings, cell biology, z-stacks and volumetric analysis, quantitative analysis workflows.
I am Lucas Thal, PhD, an imaging scientist. The work that moved programs was high-content imaging in industry: automated assays across many plate types and cell samples, in 2D and 3D, with machine-learning analysis pipelines, built so a team could make a decision on the data. Underneath it is advanced microscopy from academia: spinning disk, dSTORM, SIM and TIRF, custom analysis scripts, and a PhD in single-particle tracking and super-resolution imaging. If you generate image or cell data, that is what I am for first.
Book an intro call Five short questions, then pick a time. A first conversation about your data, no pitch.
Prefer email? [email protected]
I also do the discovery biology around it: iPSC neuron–glia disease models, assay cascades for antibody, siRNA/ASO and small-molecule programs, and the assessment an investor or BD team needs when they look at one of those programs.
Ex-biopharma, ex-PhenoVista. San Diego. No wet-lab of my own; I make yours, or your CRO’s, produce data that survives review.
Two doors
Imaging and cell-based science DOOR A · IMG · HCI
For anyone who generates image or cell data: drug developers, CROs, imaging and instrument vendors, academic cores, tool companies. This is what I am for first.
- The readout looks fine and proves nothing.
- The analysis pipeline is rerunnable only by the person who wrote it.
- The validation package has to convince your customers, not just you.
Therapeutics programs DOOR B · TX · ANY MODALITY
Discovery biology for therapeutics, in any modality — antibody, siRNA and ASO, small molecule. Neuro and immunology are the spearhead, not the fence.
- Six assays and no decision.
- The CRO will run whatever you ask for, and nobody senior is writing the ask.
- The data package has to survive a partner’s reviewer, not just your own team.
What I am for
Images that hold up. BLK 01 · IMG · HCI
High-content imaging is where good biology goes to become an unreviewable figure. I am an imaging scientist end to end, and I work the whole chain from the photon to the number in the table. The work that moved programs was high-content imaging in industry: automated assays across many plate types and cell samples, in 2D and 3D, run at screening scale and analysed so the result could carry a decision. Underneath it is advanced microscopy from academia: spinning disk, dSTORM, SIM and TIRF, with custom scripts and deep analysis. I step in at whichever step has the gap, not only at the analysis: methodology, instrument selection, immunopanels, live-imaging settings, cell biology, z-stacks and volumetric analysis, quantitative workflows.
Optics and instrumentation. I have built microscopes and run them in academic and industrial settings: widefield, spinning-disk confocal, TIRF, two-photon, super-resolution (dSTORM and SIM) and single-particle tracking, and the automated high-content systems and plate handlers that turn a method into a screen. I know where the artifacts come from because I have put the parts together.
Probes and panels. I select fluorophores on their photophysics and the optics you have rather than the catalogue: brightness (extinction coefficient times quantum yield), photostability under the illumination dose the assay will really deliver, blinking and stochastic switching for dSTORM and single-molecule work, and Stokes shift, where a long-shift dye can free a channel in a crowded panel. I design multiplex panels around spectral overlap, and have designed ligand-conjugated and bioconjugated fluorescent probes, at Lawrence Berkeley National Lab and at Vanderbilt, to label neuronal proteins for single-molecule imaging.
Methodology. Fixed ICC/IF and live-cell dynamics; sample preparation; plate layout and edge effects; controls and normalization; from dishes and chambers through 24-, 96-, 384- and 1536-well plates.
Analysis. Segmentation and feature extraction, quantitative features, clustering and multivariate statistics, machine-learning image analysis in MATLAB and Python, and single-particle tracking analysis of diffusion, clustering and membrane dynamics.
Phenotypic discovery. The readouts I have built, in 2D and 3D, carry neuro and immunology programs: ASC speck counting, synaptic puncta quantification, neurite outgrowth and health, microglia and astrocyte morphology classification, fibril uptake and aggregation, micronuclei and nuclear condensation, calcium imaging.
My PhD at Vanderbilt was single-particle tracking and super-resolution imaging of the dopamine transporter, with ten peer-reviewed publications on single-molecule imaging and fluorescent probes. My first industry job was running the high-content screening assay portfolio at PhenoVista, a CRO, across iPSC-derived and primary cells. In industry I then built the automated high-content assays and machine-learning image-analysis pipelines for neuroscience programs. Whatever you hand me, you get back a pipeline you can rerun and a figure you can defend.
Human models you can trust. BLK 02 · IPSC · TRI-CULTURE
An iPSC tri-culture that sort of works is a liability in a board deck and a data room. I design neuron, astrocyte, and microglia mono- and tri-culture models with explicit QC gates and disease-relevant readouts: pre-formed fibril and proteopathic-seed paradigms (seeding, uptake, aggregation), MEA network activity and burst analysis, synaptogenesis and synapse loss, complement deposition and microglial engulfment, DAM and reactive-astrocyte marker profiling, and viability and neurotoxicity panels from caspase to ATP. I have built neuron/astrocyte/microglia tri-culture high-content assays in industry and I know which phenotypes are disease and which are differentiation noise. I do not do immunohistochemistry (IHC), live-animal imaging or electron microscopy; when you need those, I will tell you who does.
A cascade, not a pile of assays. BLK 03 · CASCADE · GO/NO-GO
Six assays and no decision is the most common state of an early program. I build cascades — biochemical to functional to translational — with go/no-go criteria at every tier, for the modality you actually have. For antibodies: cell-based binding and blocking, internalization, FcR binding, ADCC and CDC. For siRNA and ASO: knockdown validation, uptake, and conjugate delivery in neural cells. For small molecules: FLIPR, HTRF, FP, and reporter assays through to phenotypic readouts. On the translational end: inflammasome and NLRP3 activation (ASC speck, IL-1β, caspase-1, pyroptosis), human whole-blood and PBMC assays, ex vivo stimulation PK/PD, receptor occupancy, phospho-flow, and MSD/ELISA biofluid biomarkers. I have led biology on a neuroimmunology program through hit-to-lead and contributed mechanism work to two programs that reached development-candidate nomination, one now in Phase 1. I have also sat on both sides of the CRO table, so the study plan I write is one a CRO can quote and you can judge.
Strategy that gets a program to the next gate. BLK 04 · STRATEGY · GATE
A program does not move because another assay ran. It moves because someone wrote down what the data has to show, and then held that line when the data arrived. I set go/no-go criteria per tier before the first plate goes on the reader, and I make the build-versus-outsource call assay by assay rather than by habit. I help select targets and build the internal case for them, write the RFP and score the CROs, and read a data package the way the reviewer on the other side of a partnership will read it. I have led biology on a neuroimmunology program through hit-to-lead and contributed mechanism work to two programs that reached development-candidate nomination, one now in Phase 1. When the science has to survive a board meeting, a diligence call or a data room, I write the version that does.
Neurocrine Biosciences, Senior Scientist 2021–2025 · PhenoVista Biosciences · Vanderbilt PhD, Chemical Biology · Lawrence Berkeley National Lab · Ten peer-reviewed publications · San Diego
How I work
- No lab of my own, on purpose. I design, oversee, and interpret. Your CRO or your bench runs the experiment. That keeps me independent of the result and lets you spend your wet-lab budget where it belongs.
- Fixed scope, fixed fee. Every engagement is a written proposal with a named deliverable, a date, and a price. Retainers are billed monthly in advance with a three-month minimum and thirty days’ notice after that.
- Your data stays yours. NDA first. Deliverables are assigned to you on payment. Analysis is built on open, reproducible tools (Python, CellProfiler, Jupyter) so you can rerun what I hand you without a license. Your data stays in your systems; I work in your tenant when you have one, on an encrypted machine when you don’t, and nothing you share is ever entered into a third-party AI service without your written consent.
- A wall on both sides. I hold my former employer’s confidential information with the same discipline I hold yours. Nothing from a former employer appears in my work for you; nothing from your program appears anywhere else. I decline engagements whose purpose is diligence on, or competitive intelligence about, my former employer or its partnered assets. For work near my prior areas, my agreement states in writing that I rely only on the public literature and what you provide.
- What I do not do. Immunohistochemistry (IHC), live-animal imaging, and electron microscopy. I keep a short list of people who do, and I refer without a fee — I take no commission, equity, or consideration of any kind from a CRO or vendor I might recommend to you.
- Independent business. A California LLC serving multiple clients, working from my own equipment on my own schedule. W-9 and standard terms available on request; I will sign your paper when you have counsel and offer mine when you do not.
The first step is a conversation about your data.
Five short questions, then pick a time. I will have read your answers before we start, and you will leave the call with at least one thing you can check tomorrow.