mukundvarma/kallisto_sleuth

Lightweight RNA-seq pipeline to go from fastqs to differential expression using kallisto and sleuth

RNA Seq pipeline for pseudoalignment, quantification of reads and differential expression of genes

Dependencies

0. python 3, R 3.4.4
1. Snakemake (Can be installed using conda)
2. Kallisto (Quantification via pseudoalignment) Complete installation instructions here - https://pachterlab.github.io/kallisto/download, can also be installed using conda if the bioconda channel is active
3. Sleuth (R Package)
4. Other R packages - clusterProfiler, ggplot2, dplyr, pheatmap, reshape2, org.EcK12.eg.db

Usage

snakemake --snakefile snakefile.py

To recreate output, just install dependencies, remove the outs directory and run this command.

Input files

Fastqs go inecoli/cdna

metadata.tsv goes in the home directory and should contain two columns - samples, groups. Example included

Outputs

outs/counts/ contains abundance estimations calculated using Kallisto

outs/objects/ contains a sleuth object that can be visualized in a shiny app to see QC metrics, differentially expressed genes, correlation heatmaps etc.

library(sleuth)
obj = readRDS('outs/objects/sleuth.object.RDS')
sleuth_live(obj)

This directory also contains a normalized TPM matrix (gene expression level), and a differential expression output table from sleuth's implementation of the Wald test.

outs/plots/ contains additional visualizations for correlation, volcano plots and pathway enrichment

Acknowledgements

Snakemake code heavily borrowed from github.com/slowkow and github.com/saketkc