All2

A tool for filtering variants from all2all comparison of multiple clones or single cells

Prerequisite:

1. Python 3.6 with the following packages
   1. Pandas
   2. Matplotlib
   3. Seaborn
   4. Numpy
2. Other dependencies:
   1. Samtools

Setup:

Download

   git clone https://github.com/abyzovlab/All2.git    

Downloading Samtools and reference

1. Samtools (download and install):

   wget https://sourceforge.net/projects/samtools/files/samtools/1.9/samtools-1.9.tar.bz2
   
   tar -xvf samtools-1.9.tar.bz2
   
   cd samtools-1.9/
   
   /.configure
   
   make
   

2. Reference (If you want to run the examples and is not a requirement):

   wget ftp://[email protected]/bundle/b37/human_g1k_v37_decoy.fasta.gz
   
   wget ftp://[email protected]/bundle/b37/human_g1k_v37_decoy.fasta.fai.gz
   
   gunzip human_g1k_v37_decoy.fasta.gz
   
   gunzip human_g1k_v37_decoy.fasta.fai.gz
   

How to use All2:

The All2 tool has three commands for SNVs and small INDELs:

1. score
2. call
3. matrix

And three commands for structural variants:

1. score_sv
2. call_sv
3. matrix_sv

Note: The options and outputs for SNVs, INDELs and SV are similar, however please keep the following in mind:

1. For structural variants, the only accepted tool is Manta and script doesn't include 'BND' variants
2. Since there is no variant allele frequency available for structural variants from manta the 'call_sv' option is slightly different than the 'call' option.
3. There are no variant allele frequency and mutation spectrum plots for structural variant mode and the format of the output files will be slightly different from the SNV/INDEL analyses.

Usage

python ALL2.py -h
usage:  python ALL2.py <command> [<args>]
       Three commands to use for SNVs and small INDELS:
               score --> Generates mosaic and germline scores.
               call --> Based on score cut-off generates sample level files/plots for mosaic,
                       germline mutations and plots variant allele frequency, mutation spectrum
               matrix --> Plot the mutation matrix
       Three commands to use for structural variants :
               score_sv --> Generates mosaic and germline scores.
               call_sv --> Based on score cut-off generates sample level files/plots for mosaic and 
                       germline mutations 
               matrix_sv --> Plot the mutation matrix              

All to all comparision

positional arguments:
 command     Please specify score/call/matrix as a command

optional arguments:
 -h, --help  show this help message and exit

*note

1. score:

This is the first step and the other two commands will need the output from this step.

Input

This options takes in a manifest file and output directory as its input. The manifest file should have the following format (see manifest_file.txt):

#Control	Case	Control_in_vcf	Case_in_vcf	Filename	Inclusion_region
control_1	case_1	control_in_vcf_1	case_in_vcf_1	path/to/file_1_1.vcf	path/to/file_1.bed
control_1	case_2	control_in_vcf_1	case_in_vcf_2	path/to/file_1_2.vcf	path/to/file_2.bed
....	....	....	....	....	....
control_n	case_m	control_in_vcf_n	case_in_vcf_m	path/to/file_n_m.vcf	path/to/file_m.bed

Note:

1. The columns "Control_in_vcf" and "Case_in_vcf" are optional and can be used when the "Control" and "Case" name don't match the control and case inside the vcf file.
2. Make sure to vcf files for all the combination of control and case specified in the vcf.
3. It is recommended to filter the vcf (technical artifacts) prior to running All2 to reduce the processing time. If this is not performed, script will automatically filter the noise, but might take longer.
4. The script automatically used only "PASS" variants and if you wish to use all variants use the option "-a"

Usage

python ALL2.py score -h
usage: ALL2.py [-h] -m MANIFEST_FILE -o OUTPUT_DIR

get_score

optional arguments:
  -h, --help            show this help message and exit
  -m MANIFEST_FILE, --manifest_file MANIFEST_FILE
                        Path to manifest file
  -o OUTPUT_DIR, --output_dir OUTPUT_DIR
                        Path to directory where results will be written

example:
python ALL2.py score -m manifest_file.txt -o output_example/

Output

It generates three output files:

1. "explanation_score.txt" which has the following information:

   Field_name	Explanation
   Chrom	Chromosome name
   Pos	Position of the mutation
   Ref	Reference allele
   Alt	Alternate allele
   Mosaic_score	Mosaic score for the mutation
   Germline_score	Germline score for the mutation
   Total_samples	Total number of samples in the study
   Number_of_samples_with_mutation	Number of samples/pairs where the mutation was called
   Samples_with_mutation	Name of the samples which have the mutation (comma separated values)
   VAF_of_samples_with_mutation	Variant allele frequency for the mutation for each sample (comma separated values)
   Number_of_comparision_per_sample	Number of comparison where this mutation was found in a sample(comma separated value)
   Excluded_samples	Samples excluded because of allele drop out

2. "Explanation_score_scatter.png" which shows the distribution of mutations in terms of mosaic and germline scores. This plot is helpful to determine the mosaic and germline score to use for calling mosaic and germline mutations in the next 'call' step. The size of the circle denotes the number of mutations.

   

3. "mutation_matrix.pkl" is an pickle file generated to be used for matrix visualisation in the 'matrix' option. This in no use for the user.

2. call:

This is next step after score to call the mosaic mutation based on user defined mosaic and germline score cutoff.

Input

It takes three required parameters:

1. The output directory from the 'score' command (GET_SCORE_DIRECTORY).
2. Genome reference that was used to generate the vcf files.
3. Output directory.

And five optional parameter:

1. Allele frequency cut-off for mutations (default is 35%)
2. Mosaic score cut-off (default is 0.75)
3. Germline score cut-off (default is 0.75)
4. Mosaic score cut-off for germline (default is 0.20)
5. Germline score cut-off for mosaic (default is 0.50)

Usage

 python ALL2.py call -h
usage: ALL2.py [-h] -g GET_SCORE_DIRECTORY -r REFERENCE -o OUTPUT_DIR
               [-a AF_CUTOFF] [-ms MOSAIC_SCORE_CUTOFF]
               [-gs GERMLINE_SCORE_CUTOFF]
               [-msg MOSAIC_SCORE_CUTOFF_FOR_GERMLINE]
               [-gsm GERMLINE_SCORE_CUTOFF_FOR_MOSAIC]

apply_score

optional arguments:
  -h, --help            show this help message and exit
  -g GET_SCORE_DIRECTORY, --get_score_directory GET_SCORE_DIRECTORY
                        Path to output directory of the get_score option
  -r REFERENCE, --reference REFERENCE
                        Path to reference file
  -o OUTPUT_DIR, --output_dir OUTPUT_DIR
                        Path to directory where results will be written
  -a AF_CUTOFF, --af_cutoff AF_CUTOFF
                        Allele frequency cut-off for variants (default=0.35)
  -ms MOSAIC_SCORE_CUTOFF, --mosaic_score_cutoff MOSAIC_SCORE_CUTOFF
                        Mosaic score cut-off (default=0.75)
  -gs GERMLINE_SCORE_CUTOFF, --germline_score_cutoff GERMLINE_SCORE_CUTOFF
                        Germline score cut-off (default=0.75)
  -msg MOSAIC_SCORE_CUTOFF_FOR_GERMLINE, --mosaic_score_cutoff_for_germline MOSAIC_SCORE_CUTOFF_FOR_GERMLINE
                        Mosaic score cut-off for germline mutations
                        (default=0.2)
  -gsm GERMLINE_SCORE_CUTOFF_FOR_MOSAIC, --germline_score_cutoff_for_mosaic GERMLINE_SCORE_CUTOFF_FOR_MOSAIC
                        Germline score cut-off for mosaic mutation
                        (default=0.5)

example:
python ALL2.py call -r human_g1k_v37_decoy.fasta -o output_example/ -g output_example/ -ms 0.75 -gs 0.75 

Output

This command creates a new annotated explanation score plot and three sets of folder in the output directory:

1. Annotated explanation plot:
   1. Red area is noise
   2. Green area is mosaic mutations
   3. Gray area is high frequency mosaic mutations
   4. Blue are is germline mutations

1. 'Mutation_counts'
   Two plots are generated under this:

   1. 'mutation_type_count.png', which give a mutation count for mosaic, germline and noisy mutation

      

   2. 'per_sample_mutation_count.png', which gives a sample level mutation count after filtering at the specified VAF cut-off

      

2. 'vaf_plots'
   For each sample three variant allele frequency plots are generated:

   1. Mosaic mutation

      

   2. Germline mutation

      

   3. Noise mutation

      

3. 'mutation_spectrum'
   For each sample two mutation spectrum plots are generated:
   1. 96 mutation spectrum

      1. Mosaic

         

      2. Germline

         

      3. Noise

         

   2. 6 mutation spectrum plot:

      1. Mosaic

         

      2. Germline

         

      3. Noise

         

2. matrix:

This command lets you look at the mutation matrix to verify a certain mutation

Input

It takes three parameters:

1. The output directory from the 'score' command (GET_SCORE_DIRECTORY).
2. Output directory.
3. Mutation in the format of chr_pos_alt_ref

Usage

python ALL2.py matrix -h
usage: ALL2.py [-h] -g GET_SCORE_DIRECTORY -o OUTPUT_DIR -m MUTATION

Plotting mutation matrix

optional arguments:
 -h, --help            show this help message and exit
 -g GET_SCORE_DIRECTORY, --get_score_directory GET_SCORE_DIRECTORY
                       Path to output folder of get_score
 -o OUTPUT_DIR, --output_dir OUTPUT_DIR
                       Path to directory where results will be written
 -m MUTATION, --mutation MUTATION
                       Mutation format should be chr_pos_alt_ref(eg.
                       chr1_43504477_T_C)

example:
python ALL2.py matrix -g output_example/ -o output_example/ -m chr15_80509588_A_C

Output

It creates a matrix plot along with variant allele frequency at each sample for the given mutation.