A Sparse Constrained Optimization Method for Resolving Coincident Single-Cell Events in Microfluidic-Based Impedance...

Area of Science:

• Biotechnology
• Microfluidics
• Signal Processing

Background:

• Label-free electrical impedance is crucial for single-cell analysis, including sorting and phenotyping.
• High cell concentrations cause coincident events, leading to inaccurate segmentation and distorted waveforms.
• This results in significant statistical errors in electrical phenotyping and cell counting.

Purpose of the Study:

• To develop a robust algorithm for accurate single-cell detection and analysis in high-concentration suspensions.
• To address the challenge of coincident events in electrical impedance measurements without microfluidic chip modification.
• To improve the statistical accuracy and reliability of cell counting and phenotyping.

Main Methods:

• A two-step sparse-constrained optimization algorithm utilizing $\ell _{1}$-norm regularization was proposed.
• The algorithm employs a waveform detection dictionary for signal segmentation.
• A de-coincidence dictionary is subsequently used to resolve overlapping cell waveforms.

Main Results:

• The algorithm demonstrated robust counting accuracy for particle concentrations from 2×10^5 to 5×10^6 particles/ml.
• Accuracy remained high (98.4%-99.9%) even with additive noise.
• Enhanced size discrimination and statistical accuracy were observed for polystyrene beads and T cells compared to conventional methods.

Conclusions:

• The sparse optimization method effectively segments and decomposes coincident signals into individual cell events.
• This algorithm offers a significant advancement for applications requiring precise cell counting and classification across a wide range of concentrations.
• The approach is well-suited for real-time analysis of cell/particle suspensions.