Effectively and sensitively probing the cellular phenotype and genetic heterogeneity is essential to help discover potential intrinsic regulation mechanisms in the biological system. Although biohybrid sensing techniques were widely developed, there are still lack of efficient and robust platforms for spatiotemporal cellular monitoring. The overall goal of the dissertation is to develop integrative biohybrid system for spatiotemporal cellular physical and genetic heterogeneity probing. Integrative biosensing techniques with machine learning were utilized for high-throughput profiling of single cell phenotype activity or the genetic heterogeneity. Mainly two types of biohybrid systems were developed, one is the cardiomyocytes based biohybrid TENG system for continuously temporal sensing and another one is the spectrum FISH technique for tissue wide spatial profiling of post-transcriptional regulation.

In chapter 2, we introduce a novel pharmacological screening platform that combines a biohybrid triboelectric nanogenerator (TENG) and informatic analysis for self-powered, noninvasive, and label-free biosensing in cardiac cells. The cyclic mechanical activity of functional cardiomyocytes is dynamically captured by a specially designed biohybrid TENG device and is analyzed by a custom-made machine learning algorithm to reveal distinctive fingerprints in response to different pharmacological treatment. The core of the TENG device is a multilayer mesh substrate with microscale-gapped triboelectric layers, which are induced to generate electrical outputs by the characteristic motion of cardiomyocytes upon pharmaceutical treatment. The following bioinformatic extraction from the recorded TENG signal is sufficient to predict a drug’s identity and efficacy, demonstrating the great potential of this platform as a biocompatible, low-cost, and highly sensitive drug screening system.

Further, in chapter 3, we present a universal platform for spatial profiling of diverse types of biotargets in fresh tissue via Spectrum-FISH (Spectrum digitization barcoding enhanced highly multiplex molecular FISHing). Functionalized silicon nanoprobe array works as hooks for fishing targets from fresh tissue, followed by in situ highly multiplexed hybridization via spectrum codes. The whole coding process can be finished in 2 hours with a spatial resolution of 5 μm. Spectrum-codes utilizes the permutation and combination of spectrums to break through the throughput limitation of traditional fluorescence detection (Normally 3 to 4 channels) and can achieve throughput of N×(N_ary^C-1), where N_ary indicates the mixed ratio number, C indicates the fluorophore channels number and N indicates the hybridization round. In a trinary system with 7 fluorophores with five round hybridizations, the theoretic throughput will be over 10000. Spectrum-FISH is a universal platform customized for detection of multiple types targets simply by changing the ‘bait’ in silicon nanoneedle ‘hook’. Spatial profiling of miRNAs and m⁶A were performed using the Spectrum-FISH platform in fresh olfactory bulb slices as basic demonstrations. miRNA expression-based functional clustering of brain region was performed. The correlation of miRNAs and m⁶A mRNA expression in olfactory bulb was also discussed.