Section: New Results

Optimisation

Chloroplast genome assembly

Participants : Sébastien Francois, Roumen Andonov, Dominique Lavenier.

This research focuses on the last two stages of de novo genome assembly, namely, scaffolding and gap-filling, and shows that they can be solved as part of a single optimization problem. Our approach is based on modeling genome assembly as a problem of finding a simple path in a specific graph that satisfies as many distance constraints as possible encoding the read-pair insert-size information. We formulate it as a mixed-integer linear programming (MILP) problem and apply an optimization solver to find the exact solutions on a benchmark of chloroplast genomes. We show that the presence of repetitions in the set of unitigs is the main reason for the existence of multiple equivalent solutions that are associated to alternative subpaths. We also describe two sufficient conditions and we design efficient algorithms for identifying these subpaths. Comparisons of the results achieved by our tool with the ones obtained with recent assemblers are also presented [11].

Integer Linear Programming for Metabolic Networks

Participants : Kerian Thuillier, Roumen Andonov.

Metabolic networks are a helpful tool to represent and study cell metabolisms. They contain information about every reaction occurring inside an organism. However, metabolic networks of poorly studied species are often incomplete. It is possible to complete these networks with knowledge of other well-known species.

In this study, we present a new linear programming approach for the problem of topological activation in metabolic networks based on flows and the Miller, Tucker and Zemlin (MTZ) formulation for solving the longest path problem. We developed a tool called Flutampl with ampl (A Mathematical Programming Language). It returns optimal solutions for the hybrid completion directly from sbml files (the data format used for modelling metabolic networks) [37].

Integer Linear Programming for De novo Long Reads Assembly

Participants : Victor Epain, Roumen Andonov, Dominique Lavenier.

To tackle the de novo long read assembly problem, we investigate a new 2-step method based on integer linear programming. The first step orders the long reads and the second one generates a consensus sequence. Each step is based on a different IPL specification. In 2019, we focused on step 1: long reads are first compared to build an overlapping graph. Then we use integer linear programming to find the heaviest path in a graph $G=(V,E,\lambda )$, where $V$ is the vertices set corresponding to the long reads, $E$ the edge set associated to the overlaps between long reads and $\lambda$ the overlap length. For large graph, $V$ is partitioned into several parts, each one is solved independently, and the solutions are merged together. Preliminary experimentation show that bacteria assemblies can be successfully solved in a few minutes [31].