seqQscorer

seqQscorer assesses NGS data quality by extracting statistical features from FastQ and BAM reports and applying tree-based and deep learning classifiers to predict low-quality samples.

Key Features:

• Quality Feature Analysis: Processes quality statistics and report summaries derived from FastQ and BAM files to characterize sequencing integrity.
• Machine Learning Models: Employs tree-based and deep learning classification algorithms to predict the probability that an input sample is of low quality and provides pre-trained models.
• Generalizability: Models have been validated on internal datasets and external disease diagnostic datasets and shown to generalize to previously unseen species.
• Statistical Guidelines: Derives statistical guidelines from analyses to inform automatic quality control of NGS data.
• Implementation: Implemented in Python.

Scientific Applications:

• Disease Diagnostics: Supports quality assessment of sequencing data used in disease diagnosis and personalized medicine workflows.
• Genomic Research: Maintains data quality for genetics and evolutionary biology studies.
• Data Quality Control: Enables automated quality control preprocessing to reduce errors in downstream analyses.

Methodology:

Extracts quality statistics and report summaries from FastQ and BAM files, performs statistical characterization of common NGS quality features, and applies tree-based and deep learning classifiers trained and validated on diverse internal and external disease diagnostic datasets to predict the probability of low-quality samples.