Deep learning enabled label-free microfluidic droplet classification for single cell functional assays.

Droplet-based microfluidics techniques coupled to microscopy allow for the characterization of cells at the single-cell scale. However, such techniques generate substantial amounts of data and microscopy images that must be analyzed. Droplets on these images usually need to be classified depending on the number of cells they contain. This verification, when visually carried out by the experimenter image-per-image, is time-consuming and impractical for analysis of many assays or when an assay yields many putative droplets of interest. Machine learning models have already been developed to classify cell-containing droplets within microscopy images, but not in the context of assays in which non-cellular structures are present inside the droplet in addition to cells. Here we develop a deep learning model using the neural network ResNet-50 that can be applied to functional droplet-based microfluidic assays to classify droplets according to the number of cells they contain with >90% accuracy in a very short time. This model performs high accuracy classification of droplets containing both cells with non-cellular structures and cells alone and can accommodate several different cell types, for generalization to a broader array of droplet-based microfluidics applications.

Distinct evolution of SARS-CoV-2 Omicron XBB and BA.2.86/JN.1 lineages combining increased fitness and antibody evasion.

The unceasing circulation of SARS-CoV-2 leads to the continuous emergence of novel viral sublineages. Here, we isolated and characterized XBB.1, XBB.1.5, XBB.1.9.1, XBB.1.16.1, EG.5.1.1, EG.5.1.3, XBF, BA.2.86.1 and JN.1 variants, representing >80% of circulating variants in January 2024. The XBB subvariants carry few but recurrent mutations in the spike, whereas BA.2.86.1 and JN.1 harbor >30 additional changes. These variants replicated in IGROV-1 but no longer in Vero E6 and were not markedly fusogenic. They potently infected nasal epithelial cells, with EG.5.1.3 exhibiting the highest fitness. Antivirals remained active. Neutralizing antibody (NAb) responses from vaccinees and BA.1/BA.2-infected individuals were markedly lower compared to BA.1, without major differences between variants. An XBB breakthrough infection enhanced NAb responses against both XBB and BA.2.86 variants. JN.1 displayed lower affinity to ACE2 and higher immune evasion properties compared to BA.2.86.1. Thus, while distinct, the evolutionary trajectory of these variants combines increased fitness and antibody evasion.

Distinct evolution of SARS-CoV-2 Omicron XBB and BA.2.86/JN.1 lineages combining increased fitness and antibody evasion.

The unceasing circulation of SARS-CoV-2 leads to the continuous emergence of novel viral sublineages. Here, we isolate and characterize XBB.1, XBB.1.5, XBB.1.9.1, XBB.1.16.1, EG.5.1.1, EG.5.1.3, XBF, BA.2.86.1 and JN.1 variants, representing >80% of circulating variants in January 2024. The XBB subvariants carry few but recurrent mutations in the spike, whereas BA.2.86.1 and JN.1 harbor >30 additional changes. These variants replicate in IGROV-1 but no longer in Vero E6 and are not markedly fusogenic. They potently infect nasal epithelial cells, with EG.5.1.3 exhibiting the highest fitness. Antivirals remain active. Neutralizing antibody (NAb) responses from vaccinees and BA.1/BA.2-infected individuals are markedly lower compared to BA.1, without major differences between variants. An XBB breakthrough infection enhances NAb responses against both XBB and BA.2.86 variants. JN.1 displays lower affinity to ACE2 and higher immune evasion properties compared to BA.2.86.1. Thus, while distinct, the evolutionary trajectory of these variants combines increased fitness and antibody evasion.