delanalysis 0.2.3

An analysis algoritm that is a companion to NGS-Barcode-Count

DEL-Analysis

DNA encoded library analysis. This is companion software to NGS-Barcode-Count for outputing analysis and graphs.

Table of Contents

• Installation
• Files Needed
• Run
• Methods

Installation

Anaconda python required for the instructions below

Create a del environment and activate

conda create -n del python=3.9
conda activate del

Install

From pypl:

pip install delanalysis

From source:

git clone https://github.com/Roco-scientist/DEL-Analysis.git
cd DEL-Analysis
pip install --use-feature=in-tree-build .

Files Needed

Output files from NGS-Barcode-Count

Run

Start

conda activate del
python

Working with merged data output

All code below is within python

import delanalysis

# Import merged data output from NGS-Barcode-Count.  This creates a DelDataMerged object
merged_data = delanalysis.read_merged("test_counts.all.csv")

# zscore, then quantile_normalize, then subtract background which is 'test_1'
merged_data_transformed = merged_data.binomial_zscore().subtract_background(background_name="test_1")

# Create a 2d comparison graph between 'test_2' and 'test_3' in the current directory and with a low end cutoff of 4
merged_data_transformed.comparison_graph(x_sample="test_2", y_sample="test_3", out_dir="./", min_score=4)

# Creates a DelDataSample object from a single sample from the merged object
test_2_data_transformed = merged_data_transformed.sample_data(sample_name="test_2")

# Create a 3d graph with each axis being a barcode within the current directory and a low end cutoff of 4
test_2_data_transformed.graph_3d(out_dir="./", min_score=4)

# Create a 2d graph within the current directory and a low end cutoff of 4
test_2_data_transformed.graph_2d(out_dir="./", min_score=4)

# Can all be done in one line
delanalysis.read_merged("test_counts.all.csv").binomial_zscore().subtract_background("test_1").sample_data("test_2").graph_3d("./", 4)


# Create a comparison graph for tri, di, and mono synthons
full = read_merged("../../test_del/test.all.csv")
double = read_merged("../../test_del/test.all.Double.csv")
single = read_merged("../../test_del/test.all.Single.csv")
full_double = full.concat(double)
full_double_single = full_double.concat(single)
full_double_single_zscore = full_double_single.binomial_zscore_sample_normalized()
full_double_single_zscore.subtract_background("test_1", inplace=True)
full_double_single_zscore.comparison_graph("test_2", "test_3", "../../test_del/", 0.002)

Working with sample data output

All code below is within python

import delanalysis

# Import sample data output from NGS-Barcode-Count.  This creates a DelDataSample object
sample_data = delanalysis.read_sample("test_1.csv")

# zscore
sample_data_zscore = sample_data.binomial_zscore()

# Create a 3d graph with each axis being a barcode within the current directory and a low end cutoff of 4
sample_data_zscore.graph_3d(out_dir="./", min_score=4)

# Create a 2d graph within the current directory and a low end cutoff of 4
sample_data_zscore.graph_2d(out_dir="./", min_score=4)

Resulting graphs

The actual graphs will be interactive HTML graphs with hover data etc.

From comparison_graph()

From graph_2d()

From graph_3d()

Methods

delanalysis methods to import data

Method	Description
read_merged(file_path)	Creates a DelDataMerged object which can use the methods below
read_sample(file_path)	Creates a DelDataSample object which can use the methods below

Common to merged data and sample data

Used with either delanalysis.read_merged() or delanalysis.read_sample() objects

Method	Description
building_block_columns()	returns all column names which contain building block info
data_columns()	returns all column names which contain data
data_descriptor()	Returns data_type with underscores for file output
data_type	The data type of the DelData
to_csv(out_file)	Writes the DelData object to the out_file in csv format
zscore(inplace=False)	z-scores the data
binomial_zscore(del_library_size, inplace=False)	z-scores the data using the binomial distribution standard deviation
binomial_zscore_sample_normalized(del_library_size, inplace=False)	z-scores the data using the binomial distribution standard deviation and normalizes by sqrt(n). See: Quantitative Comparison of Enrichment...
enrichment(del_library_size, inplace=False)	count * library_size/ total_counts
update_synthon_numbers(unique_synthons_per_barcode: List[int])	The number of unique synthons is inferred by the total uniques found in the data. These numbers can be updated with this function

Merged data

Used with delanalysis.read_merged() which creates a DelDataMerged object

Method	Description
quantile_normalize(inplace=False)	quantile normalizes the data
sample_enrichment(inplace=False)	(sample_count/total_sample_count)/(non_sample_count/total_non_sample_count). Still experimental as if the count only happens in one sample, a div 0 error occurs
subtract_background(background_name, inplace=False)	subtracts the background_name sample from all other samples
reduce(min_score, inplace=False)	Removes all rows from the data where no samples have a score above the min_score
merge(deldata, inplace=False)	Merges DelDataMerged data into the current DelDataMerged object
sample_data(sample_name)	Returns a DelDataSample object from the DelDataMerged object. This is needed for the 2d and 3d graph
select_samples(sample_names: List, inplace=False)	Reduces the data to the listed sample names
comparison_graph(x_sample, y_sample, out_dir, min_score=0)	Outputs a comparison graph of x_sample vs y_sample names.

Sample data

Used with delanalysis.read_sample() which creates a DelDataSample object

Method	Description
reduce(min_score, inplace=False)	reduces the data to only data greater than the min_score
max_score()	Returns the maximum score within the data
data_column()	Returns the data column name
graph_2d(out_dir, min_score=0)	Produces two subplot 2d graphs for the different barcodes of a DelDataSample.
graph_3d(out_dir, min_score=0)	Produces 3d graphs for the different barcodes of a DelDataSample.