Raman optical tweezers (ROT) as a label-free technique plays an important role in single-cell study such as heterogeneity of tumor and microbial cells. Herein we designed a chip utilizing ROT to isolate a specific single cell. The chip was made from a polydimethylsiloxane (PDMS) slab and formed into a gourd-shaped reservoir with a connected channel on a cover glass. On the chip an individual cell could be isolated from a cell crowd and then extracted with ∼0.5 μL of phosphate-buffered saline (PBS) via pipet immediately after Raman spectral measurements of the same cell. As verification, we separated four different type of cells including BGC823 gastric cancer cells, erythrocytes, lymphocytes, and E. coli cells and quantifiably characterized the heterogeneity of the cancer cells, leukocyte subtype, and erythrocyte status, respectively. The average time of identifying and isolating a specific cell was 3 min. Cell morphology comparison and viability tests showed that the successful rate of single-cell isolation was about 90%. Thus, we believe our platform could further couple other single-cell techniques such as single-cell sequencing and become a multiperspective analytical approach at the level of a single cell.

Rapid and accurate identification of individual microorganisms using single-cell Raman spectra combining with one-dimensional convolutional neural networks.

We report a new method for the rapid identification of pathogenic bacterial species at the single‐cell level that combines laser tweezers Raman spectroscopy (LTRS) with deep learning (DL). LTRS can accurately measure single‐cell Raman spectra (scRS) without destroying and labeling cells. Based on the scRS data, DL rapidly and accurately identifies pathogenic bacteria. We measured scRS of 15 species bacteria using homemade LTRS. For each species, approximately, 160 cells from three different patients were measured, one patient's data were used as test set, and the rest after being augmented was used as training set. A residual network (ResNet) model, trained on the augmented training set, achieved an accuracy of 94.53% on the test set. Moreover, we applied gradient‐weighted class activation mapping to visualize the proposed model. Finally, we demonstrated the advantages of ResNet over traditional machine‐learning algorithms.

Single-cell analysis has become a state-of-art approach to heterogeneity profiling in tumor cells. Herein, we realize a kind of single-cell multimodal analytical approach by combining single-cell RNA sequencing (scRNA-seq) with Raman optical tweezers (ROT), a label-free single-cell identification and isolation technique, and apply it to investigate drug sensitivity. The drug sensitivity of human BGC823 gastric cancer cells toward different drugs, paclitaxel and sodium dichloroacetate, was distinguished in the conjoint analytical way including morphology monitoring, Raman identification, and transcriptomic profiling. Each individual BGC823 cancer cell was measured by Raman spectroscopy, then nondestructively isolated out by ROT, and finally RNA-sequenced. Our results demonstrate each analytical mode can reflect cell response to the drugs from different perspectives and is consistent and complementary with each other. Therefore, we believe the multimodal analytical approach offers an access to comprehensive characterizations of the unicellular complexity, which especially makes sense for studying tumor heterogeneity or a desired special cell from a mixture cell sample such as whole blood.

Background Waldenström macroglobulinemia (WM) is a rare and incurable indolent B-cell malignancy. The molecular pathogenesis and the role of immunosuppressive microenvironment in WM development are still incompletely understood. Methods The multicellular ecosystem in bone marrow (BM) of WM were delineated by single-cell RNA-sequencing (scRNA-seq) and investigated the underlying molecular characteristics. Results Our data uncovered the heterogeneity of malignant cells in WM, and investigated the kinetic co-evolution of WM and immune cells, which played pivotal roles in disease development and progression. Two novel subpopulations of malignant cells, CD19+CD3+ and CD138+CD3+, co-expressing T-cell marker genes were identified at single-cell resolution. Pseudotime-ordered analysis elucidated that CD19+CD3+ malignant cells presented at an early stage of WM-B cell differentiation. Colony formation assay further identified that CD19+CD3+ malignant cells acted as potential WM precursors. Based on the findings of T cell marker aberrant expressed on WM tumor cells, we speculate the long-time activation of tumor antigen-induced immunosuppressive microenvironment that is involved in the pathogenesis of WM. Therefore, our study further investigated the possible molecular mechanism of immune cell dysfunction. A precursor exhausted CD8-T cells and functional deletion of NK cells were identified in WM, and CD47 would be a potential therapeutic target to reverse the dysfunction of immune cells. Conclusions Our study facilitates further understanding of the biological heterogeneity of tumor cells and immunosuppressive microenvironment in WM. These data may have implications for the development of novel immunotherapies, such as targeting pre-exhausted CD8-T cells in WM.