Germ-Cell Sorter
Microfluidic fluorescence-activated sorter for primordial germ cells
Studying the germline could improve infertility treatments, but the primordial germ cells (PGCs) you need come mixed with somatic cells that overgrow them, and the cells are fragile. Commercial sorters are expensive, fiddly, and rough enough on the cells to kill them. For my MEng final-year project I designed a microfluidic fluorescence-activated cell sorter (µFACS) that focused the sample gently, picked out PGCs by their GFP fluorescence, and sorted them with a pressure pulse. I built the whole thing: the fluidics, the optics, and the real-time control.
The rig
What it does
- 3D hydrodynamic flow focusing. Coaxial sheath flows squeezed the sample into a thin central core, which shrank the optical region of interest and kept shear low, since the whole point was to handle fragile cells gently.
- Simultaneous fluorescent and brightfield imaging through a telecentric relay. GFP marked the PGCs and brightfield showed every cell, so a cell that showed in brightfield but not in fluorescence was somatic, and sorting only fired on a confirmed germ cell.
- A real-time LabVIEW loop detected cells in the region of interest, then drove a pressure pump through a DAQ card to create the differential that pushed the target into the collection reservoir.
- Designed, built and tested the full instrument end to end: the chip, the optics, and the control system, with parts fabricated through the College Maker Space.
How it works
The problem
Primordial germ cells are central to understanding the germline and improving infertility treatments, but harvested samples come contaminated with somatic cells that overgrow them once differentiation factors are added. The two cell types are fragile and similar in size (roughly 10 to 20 µm), so passive methods like filters and inertial focusing can't tell them apart, and commercial FACS machines are expensive, complex, and high-shear. The job was to sort PGCs gently, cheaply, and at high purity.
How it works
The sample flowed into a 3D hydrodynamic focusing chip, where coaxial sheath flows squeezed it into a thin central core. That core was small enough to image at high magnification and gentle on fragile cells. A telecentric optical relay captured fluorescent and brightfield views at the same time: GFP fluorescence flagged the PGCs while brightfield showed every cell. A LabVIEW program read both, and when a confirmed PGC sat in the region of interest it pulsed a pressure pump (through a DAQ card) on the waste outlet to redirect that cell into the collection reservoir.
Results
Each subsystem was demonstrated as a proof of concept, not a finished sorter. Hydrodynamic focusing worked, but particle-tracking velocimetry put the focused stream about 28.5 µm off-centre and skewed, which traced back to leaking sheath connections and rough laser-cut channels, so it was not yet tight enough for real cells. The optics reliably told fluorescent particles apart from transparent ones. On the sorting side, the pressure differential redirected the central streamline and a fluorescent particle was imaged in the collection branch after an actuation, but sort purity and yield were never quantified and the collected particles were not retrieved from the reservoir. A redesigned chip, tighter focusing, and a quantified sorting test were the clear next steps.