Nested containment list
The Nested Containment List is a datastructure for interval overlap queries,
like the interval tree. It is usually an order of magnitude faster than the
interval tree both for building and query lookups.
The implementation here is a revived version of the one used in the now defunct
PyGr library, which died of bitrot. I have made it less memory-consuming and
created wrapper functions which allows batch-querying the NCLS for further speed
gains.
It was implemented to be the cornerstone of the PyRanges project, but I have made
it available to the Python community as a stand-alone library. Enjoy.
Original Paper: https://academic.oup.com/bioinformatics/article/23/11/1386/199545
Cite: http://dx.doi.org/10.1093/bioinformatics/btz615
Cite
If you use this library in published research cite
http://dx.doi.org/10.1093/bioinformatics/btz615
Install
pip install ncls
Usage
from ncls import NCLS
import pandas as pd
starts = pd.Series(range(0, 5))
ends = starts + 100
ids = starts
subject_df = pd.DataFrame({"Start": starts, "End": ends}, index=ids)
print(subject_df)
ncls = NCLS(starts.values, ends.values, ids.values)
it = ncls.find_overlap(0, 2)
for i in it:
print(i)
starts_query = pd.Series([1, 3])
ends_query = pd.Series([52, 14])
indexes_query = pd.Series([10000, 100])
query_df = pd.DataFrame({"Start": starts_query.values, "End": ends_query.values}, index=indexes_query.values)
query_df
l_idxs, r_idxs = ncls.all_overlaps_both(starts_query.values, ends_query.values, indexes_query.values)
l_idxs, r_idxs
print(query_df.loc[l_idxs])
print(subject_df.loc[r_idxs])
intervals = ncls.intervals()
intervals
There is also an experimental floating point version of the NCLS called FNCLS.
See the examples folder.
Benchmark
Test file of 100 million intervals (created by subsetting gencode gtf with replacement):
Library | Function | Time (s) | Memory (GB) |
---|
bx-python | build | 161.7 | 2.5 |
ncls | build | 3.15 | 0.5 |
bx-python | overlap | 148.4 | 4.3 |
ncls | overlap | 7.2 | 0.5 |
Building is 50 times faster and overlap queries are 20 times faster. Memory
usage is one fifth and one ninth.
Original paper
Alexander V. Alekseyenko, Christopher J. Lee; Nested Containment List (NCList): a new algorithm for accelerating interval query of genome alignment and interval databases, Bioinformatics, Volume 23, Issue 11, 1 June 2007, Pages 1386–1393, https://doi.org/10.1093/bioinformatics/btl647