Dynamic programming algorithms guarantee to find the optimal alignment between two sequences. For more than a few sequences, exact algorithms become computationally impractical, and progressive algorithms iterating pairwise alignments are widely used. These heuristic methods have a serious drawback because pairwise algorithms do not differentiate insertions from deletions and end up penalizing single insertion events multiple times. Such an unrealistically high penalty for insertions typically results in overmatching of sequences and an underestimation of the number of insertion events. We describe a modification of the traditional alignment algorithm that can distinguish insertion from deletion and avoid repeated penalization of insertions and illustrate this method with a pair hidden Markov model that uses an evolutionary scoring function. In comparison with a traditional progressive alignment method, our algorithm infers a greater number of insertion events and creates gaps that are phylogenetically consistent but spatially less concentrated. Our results suggest that some insertion/deletion “hot spots” may actually be artifacts of traditional alignment algorithms.

Genetic sequence alignment is the basis of many evolutionary and comparative studies, and errors in alignments lead to errors in the interpretation of evolutionary information in genomes. Traditional multiple sequence alignment methods disregard the phylogenetic implications of gap patterns that they create and infer systematically biased alignments with excess deletions and substitutions, too few insertions, and implausible insertion-deletion–event histories. We present a method that prevents these systematic errors by recognizing insertions and deletions as distinct evolutionary events. We show theoretically and practically that this improves the quality of sequence alignments and downstream analyses over a wide range of realistic alignment problems. These results suggest that insertions and sequence turnover are more common than is currently thought and challenge the conventional picture of sequence evolution and mechanisms of functional and structural changes.

Phylogeny-aware progressive alignment has been found to perform well in phylogenetic alignment benchmarks and to produce superior alignments for the inference of selection on codon sequences. Its implementation in the PRANK alignment program package also allows modelling of complex evolutionary processes and inference of posterior probabilities for sequence sites evolving under each distinct scenario, either simultaneously with the alignment of sequences or as a post-processing step for an existing alignment. This has led to software with many advanced features, and users may find it difficult to generate optimal alignments, visualise the full information in their alignment results, or post-process these results, e.g. by objectively selecting subsets of alignment sites.

A key component of the ongoing ENCODE project involves rigorous comparative sequence analyses for the initially targeted 1% of the human genome. Here, we present orthologous sequence generation, alignment, and evolutionary constraint analyses of 23 mammalian species for all ENCODE targets. Alignments were generated using four different methods; comparisons of these methods reveal large-scale consistency but substantial differences in terms of small genomic rearrangements, sensitivity (sequence coverage), and specificity (alignment accuracy). We describe the quantitative and qualitative trade-offs concomitant with alignment method choice and the levels of technical error that need to be accounted for in applications that require multisequence alignments. Using the generated alignments, we identified constrained regions using three different methods. While the different constraint-detecting methods are in general agreement, there are important discrepancies relating to both the underlying alignments and the specific algorithms. However, by integrating the results across the alignments and constraint-detecting methods, we produced constraint annotations that were found to be robust based on multiple independent measures. Analyses of these annotations illustrate that most classes of experimentally annotated functional elements are enriched for constrained sequences; however, large portions of each class (with the exception of protein-coding sequences) do not overlap constrained regions. The latter elements might not be under primary sequence constraint, might not be constrained across all mammals, or might have expendable molecular functions. Conversely, 40% of the constrained sequences do not overlap any of the functional elements that have been experimentally identified. Together, these findings demonstrate and quantify how many genomic functional elements await basic molecular characterization.

Motivation: Progressive algorithms are widely used heuristics for the production of alignments among multiple nucleic-acid or protein sequences. Probabilistic approaches providing measures of global and/or local reliability of individual solutions would constitute valuable developments. Results: We present here a new method for multiple sequence alignment that combines an HMM approach, a progressive alignment algorithm, and a probabilistic evolution model describing the character substitution process. Our method works by iterating pairwise alignments according to a guide tree and defining each ancestral sequence from the pairwise alignment of its child nodes, thus, progressively constructing a multiple alignment. Our method allows for the computation of each column minimum posterior probability and we show that this value correlates with the correctness of the result, hence, providing an efficient mean by which unreliably aligned columns can be filtered out from a multiple alignment. Availability: The software is freely available http://www.ulb.ac.be/sciences/ueg/ Contact: aloytyno@ulb.ac.be; mcmilink@ulb.ac.be. * To whom correspondence should be addressed.