Accelerating Minimum Common Flow Decomposition Via Safe Sequences

Presentation Type

Poster Presentation

Category

STEM (science, technology, engineering, mathematics)

Abstract/Artist Statement

Minimum Common Decomposition is a variation of the classical computer science problem Minimum Flow Decomposition. In this variant, the goal is to decompose a set of common flows into a set of weighted source-to-sink paths that satisfy the edges of each of the flows. Both flow decompositions and common decompositions serve as powerful models for multiassembly problems in bioinformatics. While flow decomposition is effective, we conjecture that common flow decomposition can more accurately recover desired sequences from highly related experiments. Previous work lends some support to this conclusion. However, experiments were limited, so further experiments are necessary to support this conclusion. Additionally, common decomposition comes with a tradeoff; it's much slower than regular flow decomposition. To remedy this, we speculate that using sequences that appear in every possible solution, safe sequences, can accelerate the decomposition process. We also intend to further generalize the idea of safe sequences into a hierarchy or hierarchies to better capture their effectiveness in solving common and other decomposition problems.

Mentor Name

Lucia Williams

This document is currently not available here.

Share

COinS
 
Mar 6th, 1:00 PM Mar 6th, 2:00 PM

Accelerating Minimum Common Flow Decomposition Via Safe Sequences

UC North Ballroom

Minimum Common Decomposition is a variation of the classical computer science problem Minimum Flow Decomposition. In this variant, the goal is to decompose a set of common flows into a set of weighted source-to-sink paths that satisfy the edges of each of the flows. Both flow decompositions and common decompositions serve as powerful models for multiassembly problems in bioinformatics. While flow decomposition is effective, we conjecture that common flow decomposition can more accurately recover desired sequences from highly related experiments. Previous work lends some support to this conclusion. However, experiments were limited, so further experiments are necessary to support this conclusion. Additionally, common decomposition comes with a tradeoff; it's much slower than regular flow decomposition. To remedy this, we speculate that using sequences that appear in every possible solution, safe sequences, can accelerate the decomposition process. We also intend to further generalize the idea of safe sequences into a hierarchy or hierarchies to better capture their effectiveness in solving common and other decomposition problems.