RESUMO
Generative pretrained models have achieved remarkable success in various domains such as language and computer vision. Specifically, the combination of large-scale diverse datasets and pretrained transformers has emerged as a promising approach for developing foundation models. Drawing parallels between language and cellular biology (in which texts comprise words; similarly, cells are defined by genes), our study probes the applicability of foundation models to advance cellular biology and genetic research. Using burgeoning single-cell sequencing data, we have constructed a foundation model for single-cell biology, scGPT, based on a generative pretrained transformer across a repository of over 33 million cells. Our findings illustrate that scGPT effectively distills critical biological insights concerning genes and cells. Through further adaptation of transfer learning, scGPT can be optimized to achieve superior performance across diverse downstream applications. This includes tasks such as cell type annotation, multi-batch integration, multi-omic integration, perturbation response prediction and gene network inference.
Assuntos
Análise de Célula Única , Análise de Célula Única/métodos , Humanos , Biologia Computacional/métodos , Inteligência Artificial , Redes Reguladoras de Genes , Algoritmos , MultiômicaRESUMO
The 3' untranslated regions (3'UTRs) of messenger RNAs contain many important cis-regulatory elements that are under functional and evolutionary constraints. It is hypothesized that these constraints are similar to grammars and syntaxes in human languages and can be modeled by advanced natural language techniques such as Transformers, which has been very effective in modeling complex protein sequence and structures. Here 3UTRBERT is described, which implements an attention-based language model, i.e., Bidirectional Encoder Representations from Transformers (BERT). 3UTRBERT is pre-trained on aggregated 3'UTR sequences of human mRNAs in a task-agnostic manner; the pre-trained model is then fine-tuned for specific downstream tasks such as identifying RBP binding sites, m6A RNA modification sites, and predicting RNA sub-cellular localizations. Benchmark results show that 3UTRBERT generally outperformed other contemporary methods in each of these tasks. More importantly, the self-attention mechanism within 3UTRBERT allows direct visualization of the semantic relationship between sequence elements and effectively identifies regions with important regulatory potential. It is expected that 3UTRBERT model can serve as the foundational tool to analyze various sequence labeling tasks within the 3'UTR fields, thus enhancing the decipherability of post-transcriptional regulatory mechanisms.
RESUMO
Hydrogen-Bonded organic frameworks (HOFs) are a type of emerging porous materials. At present, little research has been conducted on their solution state. This work demonstrates that HOFs fragment into small particles while maintaining their original assemblies upon dispersing in solvents, as confirmed by Cryo-electron microscopy coupled with 3D electron diffraction technology. 1D and 2D-Nuclear Magnetic Resonance (NMR) and zeta potential analyses indicate the HOF-based colloid solution and the isolated molecular solution have significant differences in intermolecular interactions and aggregation behavior. Such unique solution processibility allows for fabricating diverse continuous HOF membranes with high crystallinity and porosity through solution-casting approach on various substrates. Among them, HOF-BTB@AAO membranes show high C3H6 permeance (1.979 × 10-7 mol·s-1·m-2·Pa-1) and excellent separation performance toward C3H6 and C3H8 (SF = 14). This continuous membrane presents a green, low-cost, and efficient separation technology with potential applications in petroleum cracking and purification.
RESUMO
The construction of heterostructures is a universal method to hinder the radiative recombination of hot electrons and hot holes, which can effectively enhance the photothermal effect of semiconductors. In this work, a one-pot method was employed to prepare a composite named Bi2Se3@ZIF-8 NPs, which incredibly increased the photothermal conversion efficiency of Bi2Se3 NPs. The temperature elevation of Bi2Se3@ZIF-8 NPs was almost double that of the Bi2Se3 NPs; specifically, the temperature of the irradiated Bi2Se3@ZIF-8 NPs was strikingly increased to 130 °C within 6 seconds, and finally stabilized at 165 °C. Furthermore, the photothermal conversion ability was maintained over multiple irradiation cycles, which endows this composite with great potential to be an excellent photothermal agent.
RESUMO
Wnt/ß-catenin signaling dysregulation is involved in tumorigenesis. Furthermore, epigenetic modification of the Dickkopf (DKK) family (DKK14) has been shown to be important in the regulation of Wnt signaling. However, the functions and mechanism of DKK2 in the development and progression of prostate cancer remain unclear. Therefore, the present study investigated the role of DKK2 in prostate cancer. The mRNA and protein expression levels of DKK2 in prostate cancer tissues and cells were assessed by reverse transcriptionquantitative polymerase chain reaction and western blotting, respectively. The biological function of DKK2 in prostate cancer was investigated using 3(4,5dimethylthiazol2yl)-2,5diphenyltetrazolium bromide and transwell invasion assays. DKK2 was demonstrated to be upregulated in prostate cancer tissues and cells, and knockdown of DKK2 suppressed cell proliferation and invasion. Furthermore, small interfering RNA targeting DKK2 inhibited the expression of ßcatenin, cyclin D1 and cMyc in prostate cancer cells. The present report suggested that DKK2 downregulation suppressed the proliferation and invasion of prostate cancer cells by inhibiting the Wnt/ßcatenin signaling pathway.