TileSeqMut 0.1.12

Analysis scripts for TileSeq sequencing data

TileSeq mutation count package

This package is made to parse input sequecning files (fastq) with user provided parameter.csv file. Output of this pipeline is mutation counts for each pair of fastq files.

Dependencies

python 3.6/3.7/3.8 (tested mainly under py3.6)

R 3.4.4+

Bowtie2 Bowtie2-build

Installation

The alpha version is available by running:

python -m pip install TileSeqMut

If you are using conda, you can set up the envirment before installing the package:

conda install -n <env_name> pandas biopython seaborn

Then install the package cigar with pip install cigar. (Sometimes cigar is not available on with condas nor testpypi , this would need to be installed manually)

You will also need the script csv2json.R which can be installed via installing [https://github.com/jweile /tileseqMave]. Make sure csv2json.R can be found in $PATH

Execution

---

After installation, you can run the package:

tileseq_mut -p ~/path/to/paramSheet.csv -o ~/path/to/output_folder -f ~/path/to/fastq_file_folder/ -name
 name_of_the_run 

Examples:

• This command would analyze fastq files in the folder: $HOME/tileseq_data/WT/ and make a time stamped output folder with the prefix: MTHFR_test in $HOME/dev/tilseq_mutcount/output/ (Using all default parameters, see below)

# on DC
tileseq_mut -p $HOME/dev/tilseq_mutcount/190506_param_MTHFR.csv -o $HOME/dev/tilseq_mutcount/output/ -f $HOME
/tileseq_data/WT/ -name MTHFR_test

# on BC2
tileseq_mut -p $HOME/dev/tilseq_mutcount/190506_param_MTHFR.csv -o $HOME/dev/tilseq_mutcount/output/ -f $HOME
/tileseq_data/WT/ -name MTHFR_test -env BC2

Parameters

• Run tileseq_mut --help

# Required:

-p PARAM, parameter csv or json file (please see details in the input files section)
-o OUTPUT, Output directory
-f FASTQ, Path to fastq files you want to analyze (only required when you are running alignment)
-n NAME, RUN_NAME

# Optional:

-h, --help list all the args

-env ENVIRONMENT, Name of cluster you want to run this script (default = DC), you can pick from DC, BC2 or GURU.
--skip_alignment, Skip alignment for this run. Alignment output already exist and the output path should be the output generated by a previous run
-log LOG_LEVEL, Log level of the logging object: debug, info, warning, error, critical (default = debug)
-r1, R1 SAM file
-r2, R2 SAM file
-at, Alignment time required (default = 8h)
-mt, Mutation calling time required (default = 36h)
-override, This flag is used when converting the parameter sheet (csf2json). Please provide this flag if you only have one replicate.

Example of skipping alignment:

tileseq_mut -p $HOME/dev/tilseq_mutcount/190506_param_MTHFR.csv -o /home/rothlab1/rli/dev/tilseq_mutcount/output
/190506_MTHFR_WT_2020-01-29-17-07-04/ --skip_alignment

Example of running on one pair of SAM sam_files

tileseq_mut -r1 /home/rothlab1/rli//dev/tilseq_mutcount/output/MTHFR_test_2020-03-19-11-2812/sam_files/45_S45_R1_001.sam
 -r2
 /home/rothlab1/rli//dev/tilseq_mutcount/output/MTHFR_test_2020-03-19-11-28-12/sam_files/45_S45_R2_001.sam -o /home/rothlab1/rli//dev/tilseq_mutcount/output/MTHFR_test_2020-03-19-11-28-12/MTHFR_test_2020-03-19-14-32-21_mut_count -p /home/rothlab1/rli//dev/tilseq_mutcount/paramsheets/190506_MTHFR_WT.csv --skip_alignment -n MTHFR_test

Input files

---

/path/to/fastq/ - Full path to input fastq files

parameters.csv - CSV file contains information for this run (please see example here ). This file is required to be comma-seperated and saved in csv format.

Output files

---

One output folder is created for each run. The output folder are named with name_time-stamp

Within each output folder, the following files and folders will be generated:

./main.log - main logging file for alignment

./args.log - arguments for this run

./ref/ - Reference fasta file and bowtie2 index

./env_aln_sh/ - Bash scripts for submitting the alignment jobs

./sam_files/ - Alignment output and log files for the raw fastq files

./name_time-stamped_mut_count/ - Mutation counts in each sample are saved in csv files

- `./main.log` - Main log file for mutation calling

- `./args.log` - command line arguments

- `./info.csv` - Meta information for each sample: sequencing depth, tile starts/ends and # of reads mapped outside of the targeted tile

- `./count_sample_*.csv` - Raw mutation counts for each sample. With meta data in header. Variants are represented in hgvs format

- `./env_mut_sh/` - Bash scripts for summitting the mutation count jobs

- `./sample_id.log` - Log file for each sample

The count_sample_**.csv is passed to tileseqMave for further analysis

Alignment

---

The pipeline takes the sequence in the parameter file as reference and align the fastq files to the whole reference sequence. This is the sequence specified by user in the parameter file.

For each pair of fastq files (R1 and R2), the pipeline submits one alignment job to the cluster. In the folder env_sh you can find all the scripts that were submitted to the cluster when you run main.py.

Alignments were done using Bowtie2 with following parameters:

~/bowtie2 --no-head --norc --no-sq --local -x {ref} -U {r1} -S {r1_sam_file}
~/bowtie2 --no-head --nofw --no-sq --local -x {ref} -U {r2} -S {r2_sam_file}

Mutation Calls

---

From each pair of sam files we count mutations for each sample.

We first filter out reads that did not map to reference or reads that are outside of the tile. Then pass the rest of the reads to count_mut.py. Please read the wiki page about how to call mutations using CIGAR string and MD:Z tag.

In order to eliminate sequencing errors. We apply a posterior probability cut-off. The posterior probability of a mutation was calculated using the Phred scores provided in SAM files.