Stanford Tissue Mapping Center

HuBMAP补充研究提案概述 我们要求为两个项目提供补充资金： 1.开发一个基于最近开发的深度学习网络的新计算工具，称为 Orca将空间要素与Codex的单细胞表达数据一起使用，以更准确地传输 未标记的Codex数据集的像元类型注记(需要50%的资金来资助来自 斯坦福大学计算机科学的Jure Leskovec教授加入我们在斯坦福的努力 TMC)。此工具已用于将细胞类型注释传输到未标记的HuBMAP供体 小肠和大肠单细胞食典数据，已经节省了近100小时的 注记2个供体数据集所需的注记，但未将空间注记合并到 单元格类型注释的帮助。 2.用于scProtetics的EvoSep液相色谱系统(需要50%的资金)。它是世界上最大的 很少有系统可以运行真正低的纳米流速梯度(&lt；100 nL/min)。低纳米流可以 显着提高灵敏度，这是使用MASS成功的scProtetics的关键 光谱分析。

HuBMAP Supplemental Research Proposal Overview We are requesting supplemental funds for two items: 1. Development of a new computational tool based on a recently developed deep-learning network called ORCA to use spatial features with single-cell expression data from CODEX to more accurately transfer cell type annotations to unlabeled CODEX datasets (50% funds are requested to fund a member from Prof. Jure Leskovec from Computer Science at Stanford University to join our efforts at the Stanford TMC). This tool has already been used to transfer cell type annotations to unlabeled HuBMAP donor small intestine and large intestine single cell CODEX data, already saving nearly 100 hours of annotations required for annotating 2 donors datasets, yet does not incorporate spatial annotations to help with cell type annotations. 2. An EvoSep Liquid Chromatography system (50% funds are requested) for scProteomics. It is one of the few systems which can run a true low nano flow rate gradient ( <100 nl/min). Low nano flow can dramatically improve sensitivity which is the key for the success of scProteomics using mass spectrometry.