Efficient Computation of Sequence Mappability

Sequence mappability is an important task in genome re-sequencing. In the $(k,m)$-mappability problem, for a given sequence $T$ of length $n$, our goal is to compute a table whose $i$th entry is the number of indices $j \ne i$ such that length-$m$ substrings of $T$ starting at positions $i$ and $j$ have at most $k$ mismatches. Previous works on this problem focused on heuristic approaches to compute a rough approximation of the result or on the case of $k=1$. We present several efficient algorithms for the general case of the problem. Our main result is an algorithm that works in $\mathcal{O}(n \min\{m^k,\log^{k+1} n\})$ time and $\mathcal{O}(n)$ space for $k=\mathcal{O}(1)$. It requires a careful adaptation of the technique of Cole et al. [STOC 2004] to avoid multiple counting of pairs of substrings. We also show $\mathcal{O}(n^2)$-time algorithms to compute all results for a fixed $m$ and all $k=0,\ldots,m$ or a fixed $k$ and all $m=k,\ldots,n-1$. Finally we show that the $(k,m)$-mappability problem cannot be solved in strongly subquadratic time for $k,m = \Theta(\log n)$ unless the Strong Exponential Time Hypothesis fails.