CompAnno: Human-friendly comparison tools for annotated sequences.

CompAnno is a user-friendly library for comparing features across:

• genomes
• sets of genomes
• regions or subsets within the same genome

Installation

Bioconda

conda install -c induja companno

Quickstart

CompAnno uses AnnotationTable objects to load and store the contents of gff files (An annotation format provided by NCBI, and also output by prokka). Annotations can be loaded from a file as follows:

from companno import AnnotationTable, compare_counts, compare_cogs
k12 = AnnotationTable('k12.gff')
O157H7 = AnnotationTable('O157H7.gff')
HS = AnnotationTable('HS.gff')

Compare 2 genomes:

compare_counts will count the total number of annotations for each value of the given field in each AnnotationTable, and also calculate the difference between the table with the highest count and that with the lowest.

Ex. when given strand, compare_counts will count how many values are on the positive strand, and how many are on the negative strand.

compare_counts({'k12': k12, 'O157H7': O157H7}, 'strand')

Output:

	k12	O157H7	difference
+	2207	2719	512
-	2291	2727	436

Compare COGS for 2 genomes:

compare_cogs will run compare counts for the field cogs, and then annotate the output with the cog names and cog categories.

compare_cogs({'k12':k12, 'O157H7':O157H7})

Output:

	k12	O157H7	difference	category	name
COG2963	0	19	19	X	Transposase and inactivated derivatives
COG2310	0	10	10	T	Stress response protein SCP2
COG3637	1	11	10	M	Opacity protein and related surface antigens
...	...	...	...	...	...

Compare across many genomes:

compare_cogs({'k12':k12, 'O157H7':O157H7, 'HS': HS})

Output:

	k12	O157H7	HS	difference	category	name
COG2963	0	19	0	19	X	Transposase and inactivated derivatives
COG2310	0	10	0	10	T	Stress response protein SCP2
COG3637	1	11	2	10	M	Opacity protein and related surface antigens
...	...	...	...	...	...	...

Compare acrros genome sets:

k12_and_HS = AnnotationTable().append(k12).append(HS)
compare_cogs({'k12_and_HS': k12_and_HS, 'O157H7':O157H7})

Custom comparisons:

Comparison are not limited to whole genomes. Comparison functions work equally well across genome subsets:

k12_positive_strand = k12.get("strand", "+")
k12_negative_strand = k12.get("strand", "-")
compare_cogs({"+ strand" : k12_positive_strand, "- strand": k12_negative_strand})

Can use lambda functions in python to get very specific subets of an Annotation Table:

k12_20000_to_130000 = k12.get('start', lambda start: start > 20000) \
						.get('end', lambda end: end < 130000)
O157H7_20000_to_130000 = O157H7.get('start', lambda start: start > 20000) \
								.get('end', lambda end: end < 130000)
# return how many annotations are on each strand, for each subset
compare_counts({"k12_region" : k12_20000_to_130000, "O157H7_region": O157H7_20000_to_130000}, 'strand')

Under the Hood

CompAnno provides easy-to-learn wrappers around pandas, a popular C-optimized python libray for data processing.

Designed for interoperability, CompAnno AnnotationTable objects can be easily imported to and exported from pandas DataFrames.

AnnotationTable to DataFrame:

k12.as_df() # Returns pandas dataframe

DataFrame to AnnotationTable:

import pandas as pd

df = pd.read_csv('frick.csv') # where frick.csv contains annotation information using
								# gff headers (start, stop, strand, etc.)
frick = pd.DataFrame(df) ## frik is an Annotation Table