Inapproximability of maximal strip recovery

In comparative genomic, the first step of sequence analysis is usually to decompose two or more genomes into syntenic blocks that are segments of homologous chromosomes. For the reliable recovery of syntenic blocks, noise and ambiguities in the genomic maps need to be removed first. Maximal Strip Recovery (MSR) is an optimization problem proposed by Zheng, Zhu, and Sankoff for reliably recovering syntenic blocks from genomic maps in the midst of noise and ambiguities. Given $d$ genomic maps as sequences of gene markers, the objective of \msr{d} is to find $d$ subsequences, one subsequence of each genomic map, such that the total length of syntenic blocks in these subsequences is maximized. For any constant $d \ge 2$, a polynomial-time 2d-approximation for \msr{d} was previously known. In this paper, we show that for any $d \ge 2$, \msr{d} is APX-hard, even for the most basic version of the problem in which all gene markers are distinct and appear in positive orientation in each genomic map. Moreover, we provide the first explicit lower bounds on approximating \msr{d} for all $d \ge 2$. In particular, we show that \msr{d} is NP-hard to approximate within $\Omega(d/\log d)$. From the other direction, we show that the previous 2d-approximation for \msr{d} can be optimized into a polynomial-time algorithm even if $d$ is not a constant but is part of the input. We then extend our inapproximability results to several related problems including \cmsr{d}, \gapmsr{\delta}{d}, and \gapcmsr{\delta}{d}.

💡 Research Summary

The paper investigates the computational hardness of the Maximal Strip Recovery (MSR) problem, a fundamental task in comparative genomics where one seeks to extract syntenic blocks (contiguous sets of homologous genes) from multiple genomic maps that may contain noise and ambiguities. Formally, given d genomic maps, each represented as a permutation of gene markers, MSR_d asks for a subsequence from each map such that the total length of common strips (identical consecutive marker blocks) across the chosen subsequences is maximized.

Prior work had established a 2d‑approximation algorithm for constant d, but the limits of approximability were unknown. The authors first prove that MSR_d is APX‑hard for every constant d ≥ 2, even under the most restrictive setting where all markers are distinct and appear only in the positive orientation. The hardness proof proceeds via an L‑reduction from a classic APX‑hard problem such as Max‑3SAT (or Max‑E3‑LIN‑2). In this reduction, each Boolean variable and each clause are encoded as uniquely labeled marker blocks; the construction guarantees that a clause is satisfied if and only if the corresponding block can be aligned into a common strip. Because the reduction preserves approximation ratios, any PTAS for MSR_d would imply a PTAS for Max‑3SAT, contradicting known results.

Building on the reduction, the authors derive an explicit inapproximability bound of Ω(d / log d). The key observation is that the number of markers introduced per clause grows logarithmically with d, which translates the known (1 − ε) hardness of Max‑3SAT into a linear‑in‑d lower bound for MSR_d after accounting for the logarithmic blow‑up. Consequently, the previously known 2d‑approximation algorithm is essentially optimal up to a logarithmic factor; no polynomial‑time algorithm can achieve a factor better than c·d / log d for some constant c unless P = NP.

On the algorithmic side, the paper revisits the 2d‑approximation and shows that it can be implemented in polynomial time even when d is part of the input rather than a fixed constant. The authors introduce a “candidate strip” enumeration scheme that limits the search space to O(poly(n,d)) possibilities. By modeling the selection of compatible strips across the d maps as a maximum matching problem in a bipartite graph, they apply standard matching algorithms (e.g., Hopcroft‑Karp) to obtain a solution that respects the 2d guarantee. This result demonstrates that the approximation ratio does not depend on the assumption that d is constant, thereby broadening the practical applicability of the algorithm.

The hardness framework is then extended to several related formulations. For the Complementary MSR problem (CMSR_d), where the goal is to delete the minimum number of markers to achieve a set of common strips, the same L‑reduction yields APX‑hardness and the Ω(d / log d) lower bound. Moreover, the authors consider gap versions GAP‑MSR_{δ,d} and GAP‑CMSR_{δ,d}, where an additive error δ is allowed in the objective. They show that the approximation gap scales linearly with δ, implying that even relaxed versions remain hard to approximate within any factor better than O(d / log d · δ).

Overall, the paper makes three major contributions: (1) establishing APX‑hardness of MSR_d and its variants under the most basic model; (2) providing the first explicit lower bound Ω(d / log d) on the achievable approximation factor; and (3) presenting a polynomial‑time 2d‑approximation algorithm that works for arbitrary d. These results close a long‑standing gap between known algorithms and hardness, clarify the theoretical limits of syntenic block recovery, and lay a foundation for future work on tighter bounds, specialized instances (e.g., with repeated markers or reversed orientations), and empirical evaluation on real genomic data.