Installation
Minnow is written in C++14 and tested in a ubuntu server, please let us know if you have difficulty compiling it in your own machine.
git clone https://github.com/COMBINE-lab/minnow.git
cd minnow
mkdir build
cd build
cmake ..
make
Given the above steps run without error, the binary will be stored within the build/src.
How to run ?
Simple runs
Minnow has many options some are basic and some are advanced, the most basic one would require to have a
transcript to gene map (a tsv file) and the reference file in addition to that it needs three different
files to generate data from: quants_mat.csv, a gene to cell count file (comma separated), quants_mat_rows.txt,
and quants_mat_cols.txt. To list the required files, a complete file tree would look like following,
+-- reference
| +-- transcript.fasta
| +-- t2g.tsv
+-- experiment
| +-- quants_mat.csv
| +-- quants_mat_rows.txt
| +-- quants_mat_cols.txt
The t2g.tsv file contains the transcript to gene information. The csv file is the main matrix file and assumed to be
in gene to cell format and comma separated of dimension m * n where there are m genes and n cells. In the given setting
one must use a quants_mat_rows.txt with m lines and quants_mat_cols.txt with n lines. Not having this convention
might lead to failure and segmentation faults.
The matrix can be generated in many ways, like using splatter or any generative model user might want to implement. Given
the above format minnow will generate the reads keeping the matrix intact. If not specified otherwise minnow generates
100,000 molecules per cell after PCR, but this can be customized by passing the required flag with the program.
Below we will go over the options that can be provided to minnow in order to get different functionality. Minnow primarily works in two different modes. One where you can provide minnow with a simple matrix and it will with gene and cell information according to above format. Additional files can be obtained from https://github.com/COMBINE-lab/minnow
We have also linked different files in the main website.
Basic mode
In basic mode minnow takes a set of genes to assign them to whatever gene names that are provided to it. Please note that if you are using real gene names that are present in the reference then please don’t use this mode. Since in that case this mode would reassign those real gene names by it’s own algorithm and the purpose is defeated.
One the genes are assigned, minnow uses the weibull distribution to assign reads to the candidate transcripts. A typical
command in normal mode might look as follows,
src/minnow simulate --normal-mode --g2t /mnt/scratch6/avi/data/cgat/references/metadata/hg_t2g.tsv -m example_data/normal_data_100_Cells_50K_Genes/ --PCR 7 -r /mnt/scratch6/avi/data/cgat/references/txome/hg_transcriptome.fasta -e 0.01 -p 25 -o /mnt/scratch1/avi/minnow/benchmark/ismb_19_runs/splatter_test/100_50K_Genes_pbmc/defaultnorm --useWeibull
Here the folder example_data/normal_data_100_Cells_50K_Genes/ contains the matrix and the corresponding gene names
and cell names in the format shown above.
Splatter-mode
The splatter-mode is very similar to the normal-mode except, the provided matrix comes from splatter.
In either mode user can make use of the de-Bruijn graph provided in the github repo in order to generate realistic reads
in the following way
src/minnow simulate --splatter-mode --g2t /mnt/scratch7/hirak/mm/mm_t2g.tsv -m /mnt/scratch7/hirak/descend_matrices/parametric_sim/ --PCR 7 -r /mnt/scratch7/hirak/mm/mm_transcriptome_fixed.dedup.fasta -e 0.01 -p 25 -o /mnt/scratch7/hirak/descend_matrices/parametric_sim/dbg2 --dbg --gfa /mnt/scratch7/hirak/mm/mm.gfa --bfh /mnt/scratch7/hirak/mm/bfh.txt --uniq /mnt/scratch7/hirak/mm/mm_101_stranded_gene_uniqueness.txt
Alternatively a gfa file can be provided to represent de-Bruijn file.
Alevin mode
Without illumina model
src/minnow simulate --splatter-mode --g2t /mnt/scratch7/hirak/mm/mm_t2g.tsv -m /mnt/scratch7/hirak/descend_matrices/parametric_sim/ --PCR 7 -r /mnt/scratch7/hirak/mm/mm_transcriptome_fixed.dedup.fasta -e 0.01 -p 25 -o /mnt/scratch7/hirak/descend_matrices/parametric_sim/dbg2 --dbg --gfa /mnt/scratch7/hirak/mm/mm.gfa --bfh /mnt/scratch7/hirak/mm/bfh.txt --uniq /mnt/scratch7/hirak/mm/mm_101_stranded_gene_uniqueness.txt
With illumina model
src/minnow simulate --splatter-mode --g2t /mnt/scratch7/hirak/mm/mm_t2g.tsv -m /mnt/scratch7/hirak/descend_matrices/parametric_sim/ --PCR 7 -r /mnt/scratch7/hirak/mm/mm_transcriptome_fixed.dedup.fasta -e 0.01 -p 25 -o /mnt/scratch7/hirak/descend_matrices/parametric_sim/dbg2 --dbg --gfa /mnt/scratch7/hirak/mm/mm.gfa --bfh /mnt/scratch7/hirak/mm/bfh.txt --uniq /mnt/scratch7/hirak/mm/mm_101_stranded_gene_uniqueness.txt --illum ../data/illumina_error_models/ill100v5_mate2