sanjaynagi / rna-seq-pop

Snakemake workflow for Illumina RNA-sequencing experiments - extract population genomic signals from RNA-Seq data

Hello there 👋

I'm Sanjay Curtis Nagi, a researcher studying the major malaria mosquito Anopheles gambiae 🦟

rna-seq-pop's Issues

store figs in /plots folders

To make results directories neater.

possibly also store vcfs in sep folder

add venn and heatmap to documentation

and ability to use single-read data

progressiveDE - allow multiple groups comparisons

Structure R script in a way which loops through list or something, if theres more than one group to compare

get coverage script

• check number and proportion of expressed genes
• check number and proportion of of genes with SNPs called
• coverage across genome

run statement to check input files

Check for:

• fastq names match with sample names
• check chroms are in gff file
• check colnames of gene_names.tsv

Implement venn diagrams of expressed genes

• suggestion from ms reviewer

selection scan window size

#17 #15 related.

Calibrate window size of selection scans

Make video tutorial/walkthrough

Report SNP statistics

• how many genes have snps
• % nonsynonmyoys , synonymous
• transition a t g c

add CIs to Variants of interest module

Problem installing the rna-seq-pop

Hi rna-seq-pop developper,

I am having problem installing this pipeline using conda:
ls
config LICENSE README.md resources RNA-Seq-Pop-Logo.png workflow

snakemake --use-conda --cores 16
**WorkflowErrorin line 13 of /scicomp/home-pure/qlk5/Desktop/important_programs/rna-seq-pop/workflow/Snakefile:
Workflow defines configfile config/config.yaml but it is not present or accessible.
File "/scicomp/home-pure/qlk5/Desktop/important_programs/rna-seq-pop/workflow/Snakefile", line 13, in **

It looks like we need a file called config.yaml inside of the config folder , which is not exists , see bellow:
ls ../rna-seq-pop/o_config/
exampleconfig.yaml examplefastq.tsv examplesamples.tsv
__
Please, do you have any suggestion?

Cheers,
DD

automatically generate desired outputs files

write function to get desired outputs, based on config.yaml, so no editing the snakefile is necessary

Adapting workflow to other species

This looks like a comprehensive workflow. I'd like to use it to do DE analysis and variant calling for other species, like non-human malaria. What would I need to do to make that happen?

Decide what to do if multiple species regarding reference genomes

Should we allow separate species to be run in one go? also of extra importance as coluzzii and arabiensis genomes are improved (Zamyatin et al, 2020)

implement fst/pbs analysis after freebayes calling

Implement heatmap script

• A python script which produces a heatmap of normalised counts for genes of choice

Improve Volcano plot aesthetics

EnhancedVolcano R package

rename scripts

fix labels on read counts PCA

... to be the treatment (i.e Kisumu) , as opposed to sample name

look for introgression signals

rename logs

add continuous integration with travis

• Need to specify a .travis.yaml which runs snakemake on .test/
• Build set of test data in .test/

Change 'chrom' to 'contig'

currently use chrom in workflow to mean chromosome.

contig is more accurate given we work with chromosomal arms in anopheles. and will align with malariagen_data API

Check for signals.tsv

Given it's an optional step - need to check for signals file for ag1000gsweepsDE in checkINputs.py

remove need for DE contrasts rule

the rule which sets up the contrast table is unnecessary.

If its made incorrectly initially, snakemake will not delete it and it will affect future runs until the file is deleted. we could have config.yaml as an input to that rule, but it mean any change to the config.yaml will need all subsequent rules to be re-run.

ideally it will re-generate the DE.contrasts.list file each time, but not need to re-run other rules.

The alternative is to just directly use the config.yaml, which is better i think. We can have a function in tools.py which parses the config.yaml contrasts straight into a table perhaps

Fix AIMS legend position

convert shell commands to snakemake wrappers

For reproducibility, we need to convert shell commands to wrappers where they exist.

expand config.yaml

This needs to include data such as pointing to gene name info, ploidy, and ideally the option to trim sequences

complete gene set enrichment script and integrate

allowing custom snpEff databases

A new rule for creating custom snpEff databases.

change freebayes min depth param

This needs changing to 4 ideally as previous studies shown variant calling accurate above this level.
already we have missingness filters high so this probably wont make much difference.

only do FST/PBS for genes of high enough depth

Implement Anopheles gambiae AIMs chromosome plot

check that you can vary ploidy

Windowed Fst scan

apply a windowed hudson Fst scan in the same manner as PBS currently implemented

implement AIMs analysis for Anopheles gambiae datasets

(Ancestry Informative Markers)

Split regions by coverage

will make more efficient

add more options in config for everything

allow fastq files to be specified in units.tsv

Make a units.tsv file of 3 columns, sample, read1, read2 which is automatically read into snakemake.

Optionally, allow either rigid (current) or flexible filenaming by using an input function

Change syntax for contrasts

Currently, we use underscores in the config.yaml when specifying a contrast (i.e Kisumu_Tiassale), but this means that underscores cannot be used in treatment names. It's not the end of the world, but this is probably quite a regular use case.

We could use lists i.e ['Kisumu', 'Tiassale'] instead. I would prefer to keep manual specification of contrasts as it allows for the most flexibility.

It would require some minor editing of almost every script in the workflow.

Documentation

Need to fully write up methods and also a more detailed explanation of how to use pipeline and configure things

load individual packages instead of whole 'tidyverse'

dplyr/tidyr/ggplot2

name all params and snakemake@ in scripts

edit scripts to take snakemake params

• edit scripts to take snakemake params directly - aka snakemake@input[1], for R , snakemake.input[0] for python
• change rules to 'script' as opposed to 'shell'

make conda env yaml for each .smk file

So that the software can be deployed with conda, we need to specify conda envs in workflow/envs/

add some resources directives

So that we have some control over whether to run many jobs in parallel, or the same job multi-threaded

Z transform fst values

put rules into groups to be executed together

implement single-end read functionality

Currently the workflow can only handle paired-end reads. We can modify this to allow single-end data in common.smk, get_fastqs().

split the genome into freebayes jobs of roughly equal size

In order to ensure (somewhat) even running times of variant chunks, use Travis Colliers script which splits the genome based on coverage(?) in the bam index files

freebayes/freebayes#465 (comment)

this may be useful when it comes to the final region to call, which can sometimes take a lot longer than others.