Researchers use video data obtained from services like YouTube to study human behaviour; in HCI, they gain insights about the use of technology "in the wild", and help designers solve real-world problems.However, researchers wishing to conduct such studies will find little guidance in the literature to shape their own projects; methods vary, and cannot be applied universally.In this paper, we aim to help researchers decide whether an online video-based study is appropriate for their projects.Based on our own experience and surveying the literature from HCI and other fields, we discuss the benefits of studying user-uploaded videos, the proper methods to use, the inherent challenges of online video services, and the ethics of studying publicly-available video content.

This study adapts the Consensual Assessment Technique (CAT) for Large Language Models (LLMs), introducing a novel methodology for poetry evaluation.Using a 90-poem dataset with a ground truth based on publication venue, we demonstrate that this approach allows LLMs to significantly surpass the performance of non-expert human judges.Our method, which leverages forced-choice ranking within small, randomized batches, enabled Claude-3-Opus to achieve a Spearman's Rank Correlation of 0.87 with the ground truth, dramatically outperforming the best human nonexpert evaluation (SRC = 0.38).The LLM assessments also exhibited high inter-rater reliability, underscoring the methodology's robustness.These findings establish that LLMs, when guided by a comparative framework, can be effective and reliable tools for assessing poetry, paving the way for their broader application in other creative domains.

We present a mixed-methods study of professional fiction writers' experiences with generative AI (genAI), primarily focused on queer and disabled writers. Queer and disabled writers are markedly more pessimistic than others about the impact of genAI on their industry, although pessimism is the majority attitude for all groups. We explore how genAI exacerbates existing causes of precarity for writers, reasons why writers are opposed to its use, and strategies used by marginalized fiction writers to safeguard their industry.

We evaluate the amount and nature of self-disclosure in two versions of a 400-level computing ethics course focusing on discrimination and surveillance. The study involved 30 participants enrolled in two identical course offerings, taught by the same pair of instructors, but delivered in different formats: online versus in-person. Our analysis concentrated on the extent and contents of self-disclosure by both students and instructors. By using both quantitative and qualitative methods, we observed a higher prevalence of self-disclosure by both students and instructors in the online section. Notably, an analysis of demographic data revealed that minority group members were particularly active in self-disclosure in both formats. Overall, our findings suggest that an online setting may be more effective for delivering computing ethics courses where a primary goal is increasing open discussion and self-disclosure among participants.

BACKGROUND: Second-generation sequencers generate millions of relatively short, but error-prone, reads. These errors make sequence assembly and other downstream projects more challenging. Correcting these errors improves the quality of assemblies and projects which benefit from error-free reads. RESULTS: We have developed a general-purpose error corrector that corrects errors introduced by Illumina, Ion Torrent, and Roche 454 sequencing technologies and can be applied to single- or mixed-genome data. In addition to correcting substitution errors, we locate and correct insertion, deletion, and homopolymer errors while remaining sensitive to low coverage areas of sequencing projects. Using published data sets, we correct 94% of Illumina MiSeq errors, 88% of Ion Torrent PGM errors, 85% of Roche 454 GS Junior errors. Introduced errors are 20 to 70 times more rare than successfully corrected errors. Furthermore, we show that the quality of assemblies improves when reads are corrected by our software. CONCLUSIONS: Pollux is highly effective at correcting errors across platforms, and is consistently able to perform as well or better than currently available error correction software. Pollux provides general-purpose error correction and may be used in applications with or without assembly.

BACKGROUND: Illumina paired-end reads are used to analyse microbial communities by targeting amplicons of the 16S rRNA gene. Publicly available tools are needed to assemble overlapping paired-end reads while correcting mismatches and uncalled bases; many errors could be corrected to obtain higher sequence yields using quality information. RESULTS: PANDAseq assembles paired-end reads rapidly and with the correction of most errors. Uncertain error corrections come from reads with many low-quality bases identified by upstream processing. Benchmarks were done using real error masks on simulated data, a pure source template, and a pooled template of genomic DNA from known organisms. PANDAseq assembled reads more rapidly and with reduced error incorporation compared to alternative methods. CONCLUSIONS: PANDAseq rapidly assembles sequences and scales to billions of paired-end reads. Assembly of control libraries showed a 4-50% increase in the number of assembled sequences over naïve assembly with negligible loss of "good" sequence.

BACKGROUND: Identifying recombinations in HIV is important for studying the epidemiology of the virus and aids in the design of potential vaccines and treatments. The previous widely-used tool for this task uses the Viterbi algorithm in a hidden Markov model to model recombinant sequences. RESULTS: We apply a new decoding algorithm for this HMM that improves prediction accuracy. Exactly locating breakpoints is usually impossible, since different subtypes are highly conserved in some sequence regions. Our algorithm identifies these sites up to a certain error tolerance. Our new algorithm is more accurate in predicting the location of recombination breakpoints. Our implementation of the algorithm is available at http://www.cs.uwaterloo.ca/~jmtruszk/jphmm_balls.tar.gz. CONCLUSIONS: By explicitly accounting for uncertainty in breakpoint positions, our algorithm offers more reliable predictions of recombination breakpoints in HIV-1. We also document a new domain of use for our new decoding approach in HMMs.

We present a pairwise local aligner, FEAST, which uses two new techniques: a sensitive extension algorithm for identifying homologous subsequences, and a descriptive probabilistic alignment model. We also present a new procedure for training alignment parameters and apply it to the human and mouse genomes, producing a better parameter set for these sequences. Our extension algorithm identifies homologous subsequences by considering all evolutionary histories. It has higher maximum sensitivity than Viterbi extensions, and better balances specificity. We model alignments with several submodels, each with unique statistical properties, describing strongly similar and weakly similar regions of homologous DNA. Training parameters using two submodels produces superior alignments, even when we align with only the parameters from the weaker submodel. Our extension algorithm combined with our new parameter set achieves sensitivity 0.59 on synthetic tests. In contrast, LASTZ with default settings achieves sensitivity 0.35 with the same false positive rate. Using the weak submodel as parameters for LASTZ increases its sensitivity to 0.59 with high error. FEAST is available at http://monod.uwaterloo.ca/feast/.

BACKGROUND: Traditional algorithms for hidden Markov model decoding seek to maximize either the probability of a state path or the number of positions of a sequence assigned to the correct state. These algorithms provide only a single answer and in practice do not produce good results. RESULTS: We explore an alternative approach, where we efficiently compute the k paths of highest probability to explain a sequence and then either use those paths to explore alternative explanations for a sequence or to combine them into a single explanation. Our procedure uses an online pruning technique to reduce usage of primary memory. CONCLUSION: Out algorithm uses much less memory than naive approach. For membrane proteins, even simple path combination algorithms give good explanations, and if we look at the paths we are combining, we can give a sense of confidence in the explanation as well. For proteins with two topologies, the k best paths can give insight into both correct explanations of a sequence, a feature lacking from traditional algorithms in this domain.

BACKGROUND: Existing hidden Markov model decoding algorithms do not focus on approximately identifying the sequence feature boundaries. RESULTS: We give a set of algorithms to compute the conditional probability of all labellings "near" a reference labelling lambda for a sequence y for a variety of definitions of "near". In addition, we give optimization algorithms to find the best labelling for a sequence in the robust sense of having all of its feature boundaries nearly correct. Natural problems in this domain are NP-hard to optimize. For membrane proteins, our algorithms find the approximate topology of such proteins with comparable success to existing programs, while being substantially more accurate in estimating the positions of transmembrane helix boundaries. CONCLUSION: More robust HMM decoding may allow for better analysis of sequence features, in reasonable runtimes.

In 2003, Gusfield introduced the Haplotype Inference by Pure Parsimony (HIPP) problem and presented an integer program (IP) that quickly solved many simulated instances of the problem. Although it solved well on small instances, Gusfield's IP can be of exponential size in the worst case. Several authors have presented polynomial-sized IPs for the problem. In this paper, we further the work on IP approaches to HIPP. We extend the existing polynomial-sized IPs by introducing several classes of valid cuts for the IP. We also present a new polynomial-sized IP formulation that is a hybrid between two existing IP formulations and inherits many of the strengths of both. Many problems that are too complex for the exponential-sized formulations can still be solved in our new formulation in a reasonable amount of time. We provide a detailed empirical comparison of these IP formulations on both simulated and real genotype sequences. Our formulation can also be extended in a variety of ways to allow errors in the input or model the structure of the population under consideration.

BACKGROUND: Multiple genome alignment is an important problem in bioinformatics. An important subproblem used by many multiple alignment approaches is that of aligning two multiple alignments. Many popular alignment algorithms for DNA use the sum-of-pairs heuristic, where the score of a multiple alignment is the sum of its induced pairwise alignment scores. However, the biological meaning of the sum-of-pairs of pairs heuristic is not obvious. Additionally, many algorithms based on the sum-of-pairs heuristic are complicated and slow, compared to pairwise alignment algorithms. An alternative approach to aligning alignments is to first infer ancestral sequences for each alignment, and then align the two ancestral sequences. In addition to being fast, this method has a clear biological basis that takes into account the evolution implied by an underlying phylogenetic tree. In this study we explore the accuracy of aligning alignments by ancestral sequence alignment. We examine the use of both maximum likelihood and parsimony to infer ancestral sequences. Additionally, we investigate the effect on accuracy of allowing ambiguity in our ancestral sequences. RESULTS: We use synthetic sequence data that we generate by simulating evolution on a phylogenetic tree. We use two different types of phylogenetic trees: trees with a period of rapid growth followed by a period of slow growth, and trees with a period of slow growth followed by a period of rapid growth. We examine the alignment accuracy of four ancestral sequence reconstruction and alignment methods: parsimony, maximum likelihood, ambiguous parsimony, and ambiguous maximum likelihood. Additionally, we compare against the alignment accuracy of two sum-of-pairs algorithms: ClustalW and the heuristic of Ma, Zhang, and Wang. CONCLUSION: We find that allowing ambiguity in ancestral sequences does not lead to better multiple alignments. Regardless of whether we use parsimony or maximum likelihood, the success of aligning ancestral sequences containing ambiguity is very sensitive to the choice of gap open cost. Surprisingly, we find that using maximum likelihood to infer ancestral sequences results in less accurate alignments than when using parsimony to infer ancestral sequences. Finally, we find that the sum-of-pairs methods produce better alignments than all of the ancestral alignment methods.

We present a framework for improving local protein alignment algorithms. Specifically, we discuss how to extend local protein aligners to use a collection of vector seeds or ungapped alignment seeds to reduce noise hits. We model picking a set of seed models as an integer programming problem and give algorithms to choose such a set of seeds. While the problem is NP-hard, and Quasi-NP-hard to approximate to within a logarithmic factor, it can be solved easily in practice. A good set of seeds we have chosen allows four to five times fewer false positive hits, while preserving essentially identical sensitivity as BLASTP.