inst/shinyapp/ui/dashboard.md

In gongyh/RamanD2O: RamanD2O: an R/Shiny app designed for the analysis of cellular Ramanome data

title: "RamanD2O" author: "Yanhai Gong" date: "2020/12/16" output: html_document runtime: shiny

Overview

RamanD2O is a web application developed using R/Shiny to interactively analyze single cell Raman spectra data especially from cells with different metabolic activities and incorporated heavy water.

Load data

You can load your spectra data and metadata using Load data tab. For spectrum file, two columns TXT file is supported, all spectra files should be archived in a single zip file. For metadata table, it should be in TSV format with headers in the first line. The first column should be ID_Cell, feel free to add at least one columns. To try RamanD2O, you can load a small demo using the link in Load data tab.

Regular pipeline

1. Sample: subsample your dataset, optional.
2. Trim: trim wavelength, optional.
3. Filter: based on intensities or manually remove bad spectra.
4. Smooth: reduce high frequency noise, time consuming but recommended.
5. Baseline: remove baseline for each spectrum, recommended.
6. Normalize: normalize to allow equal comparison between spectrum, recommended.
7. Average: Calculate average spectra in each selected group.
8. SNR: calculate signal noise ratio (SNR) and optionally filter dataset based on SNR values (recommended).
9. CDR: calculate C-D ratio for cells incorporated heavy water.

Machine learning

1. Prepare: split the dataset into train and eval datasets.
2. Explore: visualize the dataset using t-SNE embedding (into two dimensions), optional.
3. Random forest: train a random forest classifier and evaluate for the eval dataset.

Integration analysis

1. Ramanome & Transcriptome: Integrate Ramanome and Transcriptome using O2PLS.
2. Ramanome & Metabolome: Integrate Ramanome and Metabolome, TBD.
3. MultiOmics Integration: Integrate Ramanome with multiple omics datasets, TBD.

More documents

1. Main documentation site
2. Function modules description
3. Step-by-step use case

gongyh/RamanD2O documentation built on Dec. 13, 2024, 8:39 a.m.