GenomeAnalytica

Table of Contents

1. Package Description
2. Flow
3. End to End Demo
4. User Manual
   • Supplementary Commands
   • Main Commands
5. Output
6. Contributors

Package Description

This is a package that provide tools to extract different genome architecture features, build taxonomy lineage and test features against certain value in taxonomy rank in fully automated pipeline.

The package is very user friendly allowing users to ask question like:

• Does Vertebrates have a significant GC content than other species ?
• What are the genome architecture features that significant in Eukaryotes ?
• What are the features that signifies Betacoronavirus from other Riboviria ?

Output of the genomeAnalytica package is a csv similar to common deferential expression experiments (features with adjusted p-values and averages) and a boxplot figures for significant features.

The package is providing an end to end dedicative pipeline for investigating genomic architecture features which is unique to this package as far as we know.

Flow

Prerequisites

1. Option 1 - Manual
   Download and install the latest version of the following.
   

• BASH
  bedtools
  entrez-direct
  
• Python
  pandas
  scipy
  statsmodels
  matplotlib
  seaborn
  

2. Option 2 - Automated
   Make conda environment and use the requirements.txt to install needed libraries

$ conda create --name <env_name> --file requirements.txt

$ conda activate <env_name>

End to End Demo

1. Make data directory and under it make raw directory under it make directory for each species and create a txt file containing 2 ftp links 1 for species genome gtf file and another for species genome fasta file like the following tree:
   

data
├── raw
   ├── Caenorhabditis_elegans
   │   └── WgetMe.txt
   ├── Choloepus_hoffmanni
   │   └── WgetMe.txt
   ├── Chrysolophus_pictus
   │   └── WgetMe.txt
   ├── Ciona_intestinalis
   │   └── WgetMe.txt
   ├── Ciona_savignyi
   │   └── WgetMe.txt
   ├── Drosophila_melanogaster
   │   └── WgetMe.txt
   ├── Homo_sapiens
   │   └── WgetMe.txt
   ├── Monopterus_albus
   │   └── WgetMe.txt
   ├── Mus_musculus
   │   └── WgetMe.txt
   └── Saccharomyces_cerevisiae
       └── WgetMe.txt

For using the same species as above just download from data

2. Download and unzip the data at each corresponding directory using the following command.

$ ./get_data.sh data/raw WgetMe.txt

3. Make features for all species under data/raw. Can be done either using a Fast Mode mode or Expert Mode mode

• Option 1 - Fast Mode:
  e.g: creating statistics for (all, protein coding, non protein coding) for (gene, exon, transcript, three_prime_utr, five_prime_utr)
  

$ ./make_arch_fet.sh data/raw/ g pg npg e pe npe t pt npt f pf npf h ph nph

• Option 2 - Expert Mode:
  e.g: creating statistics for protein coding exons
  

for i in `ls data/raw/`; do echo $i; ./get_len_gc.sh data/raw/$i/ exon -i transcript_biotype=protein_coding ; done

4. Make taxonomy lineage for all species under data/raw using the following command

$ ./make_tax.sh data/raw/ tax_mapping.csv

5. Make test for calculated feature under data/output using the taxonomy lineage and specifying the 10th rank (supphilum) and (vertebrata) as the value to compare against using the following command:

$ ./make_arch_test.py --stats data/output/ --tax data/tax_mapping.csv --tax_value vertebrata

User Manual

The package consists of three main commands to build standard features, build mapping and test them supplementary commands are available to batch download ftp links and build custom genome architecture features.

Supplementary Commands

1. get_data

• Description:
  This script loop throw the species directories under user specified species containing directory; reading the user specified text file containing the ftp links for gtf and fasta then unzip them
• Positional Arguments
  There are 2 mandatory arguments
  arg1: the directory containing the species directories
  arg2: the name of the files that contains the ftp link to download
  
• Example:
  

$ ./get_data.sh data/raw WgetMe.txt

2. get_len_gc

• Description
  This module gets stats (avg gc content, avg length and count ) for a custom user specified feature and filtration in a gtf file. Features extracted can differ from those served out of the box in make_arch_fet.
• Positional Arguments
  There are 2 mandatory arguments :
  arg1: the path to directory containing the species gtf and fasta files (uncompressed) with the following hirarchy: raw/species/ e.g (data/raw/Homo_Sapiens)
  arg2: the feature that will be filtered for stats. this feature should exist in column 3 gtf file. e.g (exon, gene, transcript)
  There are 2 optional arguments to be used when you want to filter certain tag in column 9 gtf:
  arg3: a flag to include or exclude the following (tag=value) argument. to include use -i , to exclude use -e
  arg4: the name of the tag you want to filter tigetherr with the filtering value in the form of (tag=value). e.g (-i gene_biotype=protein_coding)
  
• Example:
  

• to calculate the stats of all genes of Homo Sapiens use the following

$ bash get_len_gc.sh dat/raw/Homo_Sapiens gene

• to calculate the stats of exons coming from non-protein coding transcript of Homo Sapiens use the following

$ bash get_len_gc.sh dat/raw/Homo_Sapiens exon -e transcript_biotype=proteing_coding

Main Commands

3. make_arch_fet
   

• Description
  This script generalize applying stats like (avg gc content, avg length, count ) for all species under user input containing dir given that each species dir contains gtf and fasta file for species genome
• Positional Arguments
  There are minimum 2 mandatory arguments :
  arg1: the path to directory containing all the species directories containing gtf and fasta files (uncompressed). Dir must be named raw e.g (data/raw/)
  arg2,3,4,...: 1 to 3 characters representing the features that will be filtered for stats. with the following naming:
  filtration criteria: can be:
  p for pretein coding or
  np for non proteing coding
  feature type: can be:
  g for gene
  e for exon
  t for transcript
  f for five_prime_utr
  h for three_prime_utr
  e.g pg for protein coding gene; npt for non protein coding three_prime_utr
  
• Example:
  to calculate the stats of (non protein coding genes, protein coding exons and all five_prime_utr) for all species under data/raw/ use the following

$ ./ make_arch_fet.sh data/raw/ npg pe f

To calculate all features

$ ./make_arch_fet.sh data/raw/ g pg npg e pe npe t pt npt f pf npf

4. make_tax

• Description
  This script loop throw the species directories under user specified species contining directory; using the species dir naming the taxonomy lineage record is retrieved from ncbi, appended in csv file and saved in user specified file
• Positional Arguments
  There are 2 mandatory arguments
  arg1: the directory containing the species directories
  arg2: the name of the taxonomy mapping csv file to be saved
  
• Example:
  

$ ./make_tax.sh data/raw/ tax_mapping.csv

5. make_arch_test

• Description
  This script takes the path for data created by make_arch_fet, path for species lineage csv created by make_tax, rank value to test againist. Count the nulls (which is filled with 0 and returned as 0_count) in each feature, applies 2 sample 2 tailed ttest and Manwhetney, correct using benferroni and returns
  i. parsed data architicture feature in 1 csv
  ii. test results containing (feature,0_count,avg_[rank_value],avg_non_[rank_value],MannWhitney_p,ttest_p,MannWhitney_adj_p,ttest_adj_p,MannWhitney_adj_reject,ttest_adj_reject)
  iii. filtered parsed data architicture features for only significant ones according to corrected p-value < 0.05
  
• Named Arguments:
  -h, --help: show this help message and exit
  --stats [STATS]: Directory containning the nested csv stats for all species in form of (species, stats)
  --tax [TAX]: File csv contianing the full lineage of all species in stats
  --out_test [OUT_TEST]: File to write the test results
  --out_parsed [OUT_PARSED]: File to write parsed data
  --out_parsed_filtered [OUT_PARSED_FILTERED]: File to write parsed data with filtered significant features
  --out_fig [OUT_FIG] png File to write boxplot figure
  --tax_value [RANK_VALUE]: value to test for in the tax rank e.g: Vertebrata to test Vert vs In-Vert in Subphylum
  
• Example:
  

$ ./make_arch_test.py --stats data/output/ --tax data/tax_mapping.csv --tax_value vertebrata

Output

• test results

feature	0_count	avg_vertebrata	avg_non_vertebrata	MannWhitney_p	ttest_p	MannWhitney_adj_p	ttest_adj_p	MannWhitney_adj_reject	ttest_adj_reject
gene_-i_gene_biotype=protein_coding_count	0	17616.4	12821.4	0.000145036	4.61E-08	0.005221295	1.66E-06	TRUE	TRUE
exon_-e_transcript_biotype=protein_coding_avg_len	0	247.5974	382.6778	0.000145036	3.91E-05	0.005221295	0.001405841	TRUE	TRUE

• siginificant features boxplot
  

Contributors

Ahmed Tarek
Abdullah Alkhawaja
Ali El-Nisr
Montaser Bellah Yasser