Image stitching algorithm for super-resolution localization microscopy combined with fluorescence noise prior

Biomedical Optics Express, Journal Year: 2024, Volume and Issue: 15(9), P. 5411 - 5411

Published: Aug. 13, 2024

Super-resolution panoramic pathological imaging provides a powerful tool for biologists to observe the ultrastructure of samples. Localization data can maintain essential ultrastructural information biological samples with small storage space, and also new opportunity stitching super-resolution images. However, existing image methods based on localization cannot accurately calculate registration offset sample regions no or few structural points thus lead errors. Here, we proposed framework called PNanoStitcher. The fully utilizes distribution characteristics background fluorescence noise in region solves failure points. We verified our method using both simulated experimental datasets, compared it methods. PNanoStitcher achieved superior results regions. study an important driving force development digital pathology.

Language: Английский

---

Spatial multi-omics reveal intratumoral humoral immunity niches associated with tertiary lymphoid structures in pancreatic cancer immunotherapy pathologic responders

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 23, 2024

Pancreatic adenocarcinoma (PDAC) is a rapidly progressing cancer that responds poorly to immunotherapies. Intratumoral tertiary lymphoid structures (TLS) have been associated with rare long-term PDAC survivors, but the role of TLS in and their spatial relationships within context broader tumor microenvironment remain unknown. We generated multi-omics atlas encompassing 26 tumors from patients treated combination Using machine learning-enabled H&E image classification models unsupervised gene expression matrix factorization methods for transcriptomics, we characterized cellular states niches spanning across distinct morphologies Unsupervised learning TLS-specific signature significantly associates improved survival patients. These analyses demonstrate TLS-associated intratumoral B cell maturation pathological responders, confirmed proteomics BCR profiling. Our study also identifies features pathologic immune responses, revealing colocalizing IgG/IgA distribution extracellular remodeling.

Language: Английский

---

A phenotypic drug discovery approach by latent interaction in deep learning

Royal Society Open Science, Journal Year: 2024, Volume and Issue: 11(10)

Published: Oct. 1, 2024

Contemporary drug discovery paradigms rely heavily on binding assays about the bio-physicochemical processes. However, this dominant approach suffers from overlooked higher-order interactions arising intricacies of molecular mechanisms, such as those involving cis -regulatory elements. It introduces potential impairments and restrains development computational methods. To address limitation, I developed a deep learning model that leverages an end-to-end approach, relying exclusively therapeutic information drugs. By transforming textual representations virus genetic into high-dimensional latent representations, method evades challenges insufficient specificities. Its strengths lie in its ability to implicitly consider complexities epistasis chemical–genetic interactions, handle pervasive challenge data scarcity. Through various modeling skills augmentation techniques, proposed demonstrates outstanding performance out-of-sample validations, even scenarios with unknown complex interactions. Furthermore, study highlights importance chemical diversity for training. While showcases feasibility data-scarce scenarios, it reveals promising alternative situations where knowledge underlying mechanisms is limited.

Language: Английский

---

Integrative analyses of bulk, single-cell and spatial transcriptomics identified diabetes mellitus-related signature as a prognostic factor in pancreatic adenocarcinoma

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: Nov. 23, 2023

Abstract Purpose Pancreatic adenocarcinoma (PAAD) is a deadly disease, particularly for those with diabetes mellitus (DM). While there have been various studies on prognostic factors in pancreatic cancer, few specifically focused PAAD patients DM. This study aimed to identify differentially expressed genes (DEGs) between DM and non-DM individuals develop predictive model. Materials Methods were divided into training (70%) test (30%) groups, OS-associated identified using univariate COX analysis. A 10-gene risk model was constructed LASSO-penalized regression ten-fold cross-validation. Results The showed C-index of 0.83 the group 0.76 group. High represented tumor-growth angiogenic phenotype low an immune-active phenotype. Conclusion holds promise predicting overall survival DM, indicating potential benefits from immunotherapy low-risk scores.

Language: Английский

---