JARVIS

JARVIS compresses genomic sequences losslessly using competitive model prediction and arithmetic encoding to reduce storage and transmission requirements for large-scale genomic data.

Key Features:

• Lossless Compression: Ensures no data is lost during compression, preserving complete genomic sequence information.
• Competitive Prediction Model: Employs two distinct classes of models that compete to predict the best model for each symbol before encoding.
• Arithmetic Encoding: Applies arithmetic encoding to outputs of the selected predictive models.
• Weighted Context Models: Includes substitutional tolerant context models to handle sequence variations.
• Weighted Stochastic Repeat Models: Models repetitive patterns in genomic data to improve compression of repeats.
• Handling Inverted Repeats: Both model classes incorporate specialized sub-programs to process inverted repeats.

Scientific Applications:

• Genomic data storage and transmission: Reduces storage footprint and transmission bandwidth for large-scale genomic datasets.
• Anthropological analyses: Supports anthropological studies that require efficient storage or transfer of genomic sequences.
• Biomedical methods: Supports biomedical workflows that rely on lossless compressed genomic data.

Methodology:

Two classes of predictive models — weighted context models (including substitutional tolerant context models) and weighted stochastic repeat models — competitively predict the best model for each symbol, followed by arithmetic encoding; both classes include specialized sub-programs for inverted repeats.