Cell-Associated Barcode Filtering for Aggregate Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Accurate detection of cell-associated barcodes is challenging, especially in single-cell datasets with protein aggregates, leading to false molecular counts due to complications from aggregates.

Innovation Solution

A method and system for detecting aggregates by identifying suspect barcode sequences with high UMI counts, determining aggregate candidates based on detected feature thresholds, and removing aggregates from the dataset.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If aggregate sequences are included in barcode data, then the barcode can capture comprehensive genomic information, but the data quality and accuracy of single-cell sequencing results deteriorate due to sequencing bias and false positives

Engineering Contradiction:
Improvecomprehensive genomic informationVSAvoiddata quality and accuracy
Core Design Contradiction:
Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent segments barcode sequences into two distinct types: cell-associated barcodes (shorter, 8-12 bp) and aggregate-associated barcodes (longer, 14-20 bp). By length-segmenting the barcode data, the system can identify and separate true single-cell barcode signals from aggregate artifacts, thereby maintaining data quality while preserving comprehensive genomic information capture.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts and removes aggregate-associated barcode sequences from the sequencing data through computational filtering. By identifying barcodes with abnormal length characteristics and removing them, the system eliminates the harmful effect of aggregates on data quality while retaining valid single-cell sequencing data.

Inventive Principle:
Principle #2Taking out (Extraction)

2Productivity

If standard barcode sequencing methods are used, then the sequencing process is simple and fast, but aggregate sequences cause sequencing bias and reduce reliability

Engineering Contradiction:
Improvesequencing speedVSAvoidsequencing reliability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent performs preliminary identification of aggregate-associated barcodes based on their characteristic longer length (14-20 bp) before the main sequencing analysis. This preliminary filtering action is performed on the barcode data itself, allowing rapid identification and removal of aggregate sequences without requiring complex additional sequencing steps, thus maintaining high sequencing speed while improving reliability.

Inventive Principle:
Principle #10Preliminary action

3Difficulty of detecting and measuring

If longer barcode sequences are used to represent aggregates, then aggregate detection becomes possible, but the complexity of data processing increases

Engineering Contradiction:
Improveaggregate detection capabilityVSAvoiddata processing complexity
Core Design Contradiction:
Difficulty of detecting and measuringVSDevice complexity

Solution Approach 1:

The patent applies local quality differentiation by assigning different length characteristics to different types of barcodes: shorter lengths (8-12 bp) for cell-associated barcodes and longer lengths (14-20 bp) for aggregate-associated barcodes. This local differentiation in barcode length creates an easily distinguishable feature that simplifies aggregate detection and processing complexity.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4186060B1Systems and methods for detecting and removing aggregates for calling cell-associated barcodes
Publication Date: 2026.04.29 10X GENOMICS INC
  • EP4186060B1 patent drawingFigure 1
  • EP4186060B1 patent drawingFigure 2
  • EP4186060B1 patent drawingFigure 3A

AI summary

Methods and systems for removing aggregates from a dataset can be provided. For example, method for removing aggregates can comprise: receiving a first dataset comprising a plurality of sequence reads; grouping the plurality of sequence reads into bins, wherein each bin comprises sequence reads that share a common barcode sequence; identifying a subset of barcode sequences from the bins as aggregates by tracking correction events of sequence reads; removing the subset of barcode sequences from the first dataset to obtain a second dataset of sequence reads.