Machine Boss

Machine Boss constructs and manipulates bioinformatic finite-state automata and Hidden Markov Models (HMMs) to enable rapid prototyping, combination, and analysis of state-machine models.

Key Features:

• Automated State Machine Construction: Automates the construction of bioinformatic state machines to reduce manual assembly of HMMs and related automata.
• Comprehensive Operations Suite: Provides operations for concatenation, composition, intersection, reverse complementing, and Kleene-starring of automata.
• Inference and Parameter-Fitting Algorithms: Supports Forward, Backward, Viterbi, Expectation-Maximization (EM), beam search, and prefix search for inference and parameter fitting.
• One-Line Prototyping: Enables one-line "recipes" to prototype common HMMs for tasks such as protein-to-DNA alignment, DNA data storage, and nanopore sequence analysis.

Scientific Applications:

• Protein-to-DNA Alignment: Constructs state machines that model protein-to-DNA alignment processes for sequence comparison and alignment tasks.
• DNA Data Storage: Models encoding schemes for DNA data storage using HMM-based automata to represent complex sequence transformations.
• Nanopore Sequence Analysis: Builds and combines automata to support analysis workflows for nanopore sequencing data.

Methodology:

Performs automata operations (concatenation, composition, intersection, reverse complementing, Kleene-starring) and applies inference and parameter-fitting algorithms including Forward, Backward, Viterbi, Expectation-Maximization (EM), beam search, and prefix search.