mitea-hires 0.0.18

Infer single-cell and spatial microRNA activity from transcriptomics data

Welcome to the miTEA-HiRes package!

miTEA-HiRes is an open-source package, designed to easily compute high resolution microRNA activity maps. This package is used in our paper "Inferring single-cell and spatial microRNA activity from transcriptomics data" by Herbst et al [1].

If you use miTEA-HiRes in your research, please cite [1]. *** add link and publication

Installation

You will need Python 3.8 or above.

pip install mitea_hires

Support and bug report:

This package is provided by Yakhini research group. If you encounter any issues, please contact efiherbst through gmail.

Example 1 - Single-cell data with 'Total' activity mode

In this case, miTEA-HiRes computes the activity for each cell and microRNA, and produces activity maps on a UMAP layout for the most active microRNAs.

Input format:

One or more raw counts matrices should be available in the data path, in the form of zipped or unzipped 'txt', 'tsv', 'mtx' or 'pkl' files.

YOUR_DATA_FOLDER
|   counts_1.txt
│   counts_2.txt
|   ...

If 10X files (i.e. 'mtx') are used for single-cell data, your data folder should look as follows:

YOUR_DATA_FOLDER
|   *_barcodes.tsv.gz   
│   *_genes.tsv.gz
│   *.mtx.gz

Usage:

You may use the package via command-line tool:

mitea-hires --data_path='PATH_TO_YOUR_DATA' --dataset_name='DATASET_NAME'

Or by importing the library into your code:

import mitea_hires

data_path = 'PATH_TO_YOUR_DATA'
dataset_name = 'NAME_OF_YOUR_DATASET' 

mitea_hires.compute(data_path=data_path, dataset_name=dataset_name)

Example 2 - Single-cell data with 'Comaprative' activity mode

In this case, miTEA-HiRes computes microRNA differential activity for two populations of interest. microRNA activity for each population is presented on histogram and UMAP layouts for miroRNAs of potential interest.

Input format:

Same as in Example 1, with the following required preprocessing: A unique population string should be included withing each cell id string. Taking for example one cell from 'CONTROL' population and one from 'DISEASE':

1. Cell string 'AACAATGTGCTCCGAG' should be transformed to 'AACAATGTGCTCCGAG_CONTROL'.
2. Cell string 'AACAGCCTCCTGACTA' should be transformed to 'AACAGCCTCCTGACTA_DISEASE'.

Usage:

Example using command line:

mitea-hires --data_path='PATH_TO_YOUR_DATA' --dataset_name='DATASET_NAME' --populations='DISEASE','CONTROL'

Example 3 - Spatial trascriptomics data

In this case, miTEA-HiRes computes the activity for each spot and microRNA, and produces spatial activity maps for microRNAs which are most abundantly active throughout the spots.

Input format:

Make sure you have obtained the following files for your Visium data (these can be downloaded from Visium website):

filtered_feature_bc_matrix

spatial

Your data path should then look as follows:

YOUR_DATA_FOLDER
└───filtered_feature_bc_matrix
│   │   barcodes.tsv.gz
│   │   features.tsv.gz
│   │   matrix.mtx.gz
└───spatial
   │   tissue_positions_list.csv
   │   ...

Usage:

mitea-hires --data_path='PATH_TO_YOUR_DATA' --dataset_name='DATASET_NAME'

Outputs

'results' folder is generated under the provided data_path unless specified otherwise and contains:

1. csv files with the activity p-values (and other related statistics) for every cell/spot and every microRNA.
2. html file with sorted microRNAs according to their potential interest, including links to activity maps.
3. folder with activity map plots.

Recommended setup

mitea-hires is desinged to utilize all CPUs available for parallel computing, hence in order to speed up the processing time, you may want to consider using resources with more CPUs. For example, an input of spatial trascriptomics data including 2,264 spots, ~6,000 genes per spot, computing activity for 706 microRNAs using a cloud instance with 16 cores takes 22 minutes.

Advanced usage

run mitea-hires --helpfull in command line to see additional supported flags.

mitea-hires can also be imported within your python script and then, end-to-end compuation can be executed calling the 'compute' function. Alternatively, in order to use parts of the computation, other functions can be called.

Supported species

miTEA-HiRes currently supports mouse and human.

[1] Inferring single-cell and spatial microRNA activity from transcriptomics data. Herbst et al.