ionique: Nanopore Signal Analysis
Version: 0.4.0
ionique is a Python framework for processing ionic current data from nanopore experiments. It provides a unified workflow for loading raw traces, filtering noise, detecting translocation events, and extracting features — all built around a flexible tree-based segment hierarchy.
from ionique.io import EDHReader
from ionique.datatypes import TraceFile
from ionique.parsers import AutoSquareParser, SpeedyStatSplit
from ionique.utils import Filter, extract_features
# Load a nanopore recording
metadata, current, voltage = EDHReader("experiment.edh", voltage_compress=True)
trace = TraceFile(current, voltage=voltage, metadata=metadata)
# Filter noise
filt = Filter(cutoff_frequency=5000, filter_type="lowpass", sampling_frequency=100000)
filt(trace.current)
# Detect blockade events in each voltage step
detector = AutoSquareParser(threshold_baseline=0.7, expected_conductance=1.9)
trace.parse(detector, newrank="event", at_child_rank="vstep")
# Segment sub-states within each event
splitter = SpeedyStatSplit(sampling_freq=100000, min_width=50)
trace.parse(splitter, newrank="state", at_child_rank="event")
# Extract features to a DataFrame
df = extract_features(trace, "event", ["mean", "std", "duration"])
Key capabilities:
File I/O — Read
.edh,.opt, and.abfformats with automatic unit scaling and voltage-step detection.Signal preprocessing — Lowpass/highpass/bandpass filtering, clock-tone removal, and edge trimming.
Event detection — Parsers for finding translocation events: spike detection (SpikeParser), square-pulse analysis (AutoSquareParser), and threshold-based rules (lambda_event_parser).
Sub-state segmentation — SpeedyStatSplit resolves multi-level current states within detected events using variance-based recursive splitting.
Segment tree — Hierarchical data model (file → voltage step → event) with recursive parsing, feature lookup, and memory-efficient MetaSegments.
Feature extraction — Export segment statistics to pandas DataFrames with custom computed columns.
Visualization — Quick trace plotting with
qp_trace()and interactive Panel/Bokeh dashboards.
Getting Started