Minimal Conflicting Sets for the Consecutive Ones Property in ancestral genome reconstruction

A binary matrix has the Consecutive Ones Property (C1P) if its columns can be ordered in such a way that all 1’s on each row are consecutive. A Minimal Conflicting Set is a set of rows that does not have the C1P, but every proper subset has the C1P. Such submatrices have been considered in comparative genomics applications, but very little is known about their combinatorial structure and efficient algorithms to compute them. We first describe an algorithm that detects rows that belong to Minimal Conflicting Sets. This algorithm has a polynomial time complexity when the number of 1’s in each row of the considered matrix is bounded by a constant. Next, we show that the problem of computing all Minimal Conflicting Sets can be reduced to the joint generation of all minimal true clauses and maximal false clauses for some monotone boolean function. We use these methods on simulated data related to ancestral genome reconstruction to show that computing Minimal Conflicting Set is useful in discriminating between true positive and false positive ancestral syntenies. We also study a dataset of yeast genomes and address the reliability of an ancestral genome proposal of the Saccahromycetaceae yeasts.

💡 Research Summary

The paper addresses a fundamental problem in comparative genomics: how to identify the minimal subsets of rows in a binary matrix that prevent the matrix from having the Consecutive Ones Property (C1P). A matrix has the C1P if its columns can be permuted so that, in every row, the 1‑entries appear in a single contiguous block. Violations of C1P arise frequently in real genomic data because of sequencing errors, mis‑assemblies, or evolutionary rearrangements, and distinguishing true ancestral synteny blocks from artefacts is a central challenge in reconstructing ancestral genomes.

The authors introduce the notion of a Minimal Conflicting Set (MCS). An MCS is a set of rows that together do not satisfy the C1P, yet every proper subset of those rows does satisfy the C1P. Consequently, each row in an MCS is indispensable for the conflict; removing any one row restores the property. This definition refines the coarse “conflict” concept used in earlier work and provides a precise combinatorial object that pinpoints the exact rows responsible for a C1P violation.

The first technical contribution is an algorithm that detects rows that belong to at least one MCS when the number of 1’s per row is bounded by a constant (k). The algorithm models each row as a set of column positions containing 1’s and, for each row, temporarily removes it and checks whether the remaining matrix satisfies C1P. Because the number of possible column orderings grows only polynomially when each row contains at most (k) ones, the entire detection procedure runs in (O(n \cdot \text{poly}(m))) time, where (n) is the number of rows and (m) the number of columns. Thus, for typical genomic matrices where each synteny block involves a small number of markers, the algorithm is practically linear‑time.

The second, more general contribution tackles the enumeration of all MCSs. The authors observe that the C1P violation can be expressed as a monotone Boolean function (f) over binary variables representing the presence of rows. A subset of rows makes (f) true exactly when the induced submatrix lacks C1P. In this formulation, an MCS corresponds to a minimal true clause (MTC) of (f), while a maximal set of rows that still satisfies C1P corresponds to a maximal false clause (MFC). By jointly generating all MTCs and MFCs, one obtains the complete collection of MCSs without redundancy. The paper adapts classic algorithms for minimal clause generation (e.g., Berge’s algorithm) and maximal clause generation, integrating them into a unified framework that simultaneously explores the lattice of true and false assignments. Although the worst‑case complexity remains exponential—as is inevitable for exhaustive enumeration—the approach is efficient on the sparse conflict structures typical of genomic data.

To validate the methodology, the authors conduct two sets of experiments. In the first, they simulate ancestral genome scenarios, deliberately inserting false synteny blocks and random noise. The detection algorithm identifies rows involved in conflicts with high precision, and after filtering out rows belonging to any MCS, the recall of genuine ancestral syntenies improves dramatically. This demonstrates that MCSs serve as reliable indicators of erroneous data.

The second experiment applies the full MCS enumeration pipeline to a real dataset of Saccharomycetaceae yeast genomes. The authors compare their results with a previously published ancestral genome reconstruction for this clade. Several genomic intervals repeatedly appear in MCSs, suggesting that those intervals are likely mis‑ordered or contain spurious adjacencies in the proposed ancestral model. By flagging these intervals, the authors provide a concrete diagnostic tool for assessing the credibility of ancestral reconstructions and guiding manual curation or re‑analysis.

Overall, the paper makes three key contributions: (1) a polynomial‑time algorithm for detecting rows that can belong to an MCS under a bounded‑density assumption; (2) a reduction of the exhaustive MCS enumeration problem to the joint generation of minimal true and maximal false clauses of a monotone Boolean function; and (3) empirical evidence that MCS analysis can discriminate true positive from false positive ancestral syntenies in both simulated and real yeast data. The work opens several avenues for future research, including extending the detection algorithm to rows with unbounded numbers of 1’s, applying the MCS framework to other combinatorial biology problems (e.g., transcriptome assembly, protein interaction networks), and integrating MCS‑based confidence scores into automated ancestral genome reconstruction pipelines.