Dan DeBlasio

Epigenetic regulation of microRNA-375 and its role in melanoma development in humans

To identify epigenetically regulated miRNAs in melanoma, we treated a stage 3 melanoma cell line WM1552C, with 5AzadC and/or 4‐PBA. Several hypermethylated miRNAs were detected, one of which, miR‐375, was highly methylated and was studied further. Minimal CpG island methylation was observed in melanocytes, keratinocytes, normal skin, and nevus but hypermethylation was observed in patient tissue samples from primary, regional, distant, and nodular metastatic melanoma. Ectopic expression of miR‐375 inhibited melanoma cell proliferation, invasion, and cell motility, and induced cell shape changes, strongly suggesting that miR‐375 may have an important function in the development and progression of human melanomas.

Highlights from the tenth ISCB Student Council Symposium 2014

This report summarizes the scientific content and activities of the annual symposium organized by the Student Council of the International Society for Computational Biology (ISCB), held in conjunction with the Intelligent Systems for Molecular Biology (ISMB) conference in Boston, USA, on July 11th, 2014.

Parameter advising for multiple sequence alignment

Background While the multiple sequence alignment output by an aligner strongly depends on the parameter values used for its alignment scoring function (i.e. choice of gap penalties and substitution scores), most users rely on the single default parameter setting. A different parameter setting, however, might yield a much higher-quality alignment for a specific set of input sequences. The problem of picking a good choice of parameter values for a given set of input sequences is called parameter advising. A parameter advisor has two ingredients: (i) a set of parameter choices to select from, and (ii) an estimator that estimates the accuracy of a computed alignment; the parameter advisor then picks the parameter choice from the set whose resulting alignment has highest estimated accuracy. Our estimator Facet (Feature-based Accuracy Estimator) is a linear combination of real-valued feature functions of an alignment. We assume the feature functions are given as well as the universe of parameter choices from which the advisor’s set is drawn. For this scenario we define the problem of learning an optimal advisor by finding the best possible parameter set for a collection of training data of reference alignments. Learning optimal advisor sets is NP-complete [1]. For the advisor sets

Learning parameter sets for alignment advising

While the multiple sequence alignment output by an aligner strongly depends on the parameter values used for the alignment scoring function (such as the choice of gap penalties and substitution scores), most users rely on the single default parameter setting provided by the aligner. A different parameter setting, however, might yield a much higher-quality alignment for the specific set of input sequences. The problem of picking a good choice of parameter values for specific input sequences is called parameter advising. A parameter advisor has two ingredients: (i) a set of parameter choices to select from, and (ii) an estimator that provides an estimate of the accuracy of the alignment computed by the aligner using a parameter choice. The parameter advisor picks the parameter choice from the set whose resulting alignment has highest estimated accuracy. We consider for the first time the problem of learning the optimal set of parameter choices for a parameter advisor that uses a given accuracy estimator. The optimal set is one that maximizes the expected true accuracy of the resulting parameter advisor, averaged over a collection of training data. While we prove that learning an optimal set for an advisor is NP-complete, we show there is a natural approximation algorithm for this problem, and prove a tight bound on its approximation ratio. Experiments with an implementation of this approximation algorithm on biological benchmarks, using various accuracy estimators from the literature, show it finds sets for advisors that are surprisingly close to optimal. Furthermore, the resulting parameter advisors are significantly more accurate in practice than simply aligning with a single default parameter choice.

Learning Parameter-Advising Sets for Multiple Sequence Alignment

While the multiple sequence alignment output by an aligner strongly depends on the parameter values used for the alignment scoring function (such as the choice of gap penalties and substitution scores), most users rely on the single default parameter setting provided by the aligner. A different parameter setting, however, might yield a much higher-quality alignment for the specific set of input sequences. The problem of picking a good choice of parameter values for specific input sequences is called parameter advising. A parameter advisor has two ingredients: (i) a set of parameter choices to select from, and (ii) an estimator that provides an estimate of the accuracy of the alignment computed by the aligner using a parameter choice. The parameter advisor picks the parameter choice from the set whose resulting alignment has highest estimated accuracy. In this paper, we consider for the first time the problem of learning the optimal set of parameter choices for a parameter advisor that uses a given accuracy estimator. The optimal set is one that maximizes the expected true accuracy of the resulting parameter advisor, averaged over a collection of training data. While we prove that learning an optimal set for an advisor is NP-complete, we show there is a natural approximation algorithm for this problem, and prove a tight bound on its approximation ratio. Experiments with an implementation of this approximation algorithm on biological benchmarks, using various accuracy estimators from the literature, show it finds sets for advisors that are surprisingly close to optimal. Furthermore, the resulting parameter advisors are significantly more accurate in practice than simply aligning with a single default parameter choice.

SICLE: a high-throughput tool for extracting evolutionary relationships from phylogenetic trees

We present the phylogeny analysis software SICLE (Sister Clade Extractor), an easy-to-use, high-throughput tool to describe the nearest neighbors to a node of interest in a phylogenetic tree as well as the support value for the relationship. The application is a command line utility that can be embedded into a phylogenetic analysis pipeline or can be used as a subroutine within another C++ program. As a test case, we applied this new tool to the published phylome of Salinibacter ruber, a species of halophilic Bacteriodetes, identifying 13 unique sister relationships to S. ruber across the 4,589 gene phylogenies. S. ruber grouped with bacteria, most often other Bacteriodetes, in the majority of phylogenies, but 91 phylogenies showed a branch-supported sister association between S. ruber and Archaea, an evolutionarily intriguing relationship indicative of horizontal gene transfer. This test case demonstrates how SICLE makes it possible to summarize the phylogenetic information produced by automated phylogenetic pipelines to rapidly identify and quantify the possible evolutionary relationships that merit further investigation. SICLE is available for free for noncommercial use at http://eebweb.arizona.edu/sicle/.

PMFastR: a new approach to multiple RNA structure alignment

Multiple RNA structure alignment is particularly challenging because covarying mutations make sequence information alone insufficient. Many existing tools for multiple RNA alignments first generate pairwise RNA structure alignments and then build the multiple alignment using only the sequence information. Here we present PMFastR, an algorithm which iteratively uses a sequence-structure alignment procedure to build a multiple RNA structure alignment. PMFastR has low memory consumption allowing for the alignment of large sequences such as 16S and 23S rRNA. The algorithm also provides a method to utilize a multi-core environment. Finally, we present results on benchmark data sets from BRAliBase, which shows PMFastR outperforms other state-of-the-art programs. Furthermore, we regenerate 607 Rfam seed alignments and show that our automated process creates similar multiple alignments to the manually-curated Rfam seed alignments.

Highlights of the second ISCB Student Council Symposium in Africa, 2017

Student Council Symposiums (SCSs) have been found to be very useful for students and young researchers. This is especially true given that the events are held directly before large international conferences, giving attendees a chance to gain exposure and have a warm up to the social nuances involved in attending such a meeting. This was the second SCS held in Africa in conjunction with the International Society for Computational Biology (ISCB) and the African Society for Bioinformatics and Computational Biology’s (ASBCB) biennial meeting. This symposium was organised by students within the society inside Africa and was held on the 10 th of October 2017 in Entebbe, Uganda.

Predicting Core Columns of Protein Multiple Sequence Alignments for Improved Parameter Advising

In a computed protein multiple sequence alignment, the coreness of a column is the fraction of its substitutions that are in so-called core columns of the gold-standard reference alignment of its proteins. In benchmark suites of protein reference alignments, the core columns of the reference are those that can be confidently labeled as correct, usually due to all residues in the column being sufficiently close in the spatial superposition of the folded three-dimensional structures of the proteins. When computing a protein multiple sequence alignment in practice, a reference alignment is not known, so its coreness can only be predicted.

A Memory Efficient Method for Structure-Based RNA Multiple Alignment

Structure-based RNA multiple alignment is particularly challenging because covarying mutations make sequence information alone insufficient. Existing tools for RNA multiple alignment first generate pairwise RNA structure alignments and then build the multiple alignment using only sequence information. Here we present PMFastR, an algorithm which iteratively uses a sequence-structure alignment procedure to build a structure-based RNA multiple alignment from one sequence with known structure and a database of sequences from the same family. PMFastR also has low memory consumption allowing for the alignment of large sequences such as 16S and 23S rRNA. The algorithm also provides a method to utilize a multicore environment. We present results on benchmark data sets from BRAliBase, which shows PMFastR performs comparably to other state-of-the-art programs. Finally, we regenerate 607 Rfam seed alignments and show that our automated process creates multiple alignments similar to the manually curated Rfam seed alignments. Thus, the techniques presented in this paper allow for the generation of multiple alignments using sequence-structure guidance, while limiting memory consumption. As a result, multiple alignments of long RNA sequences, such as 16S and 23S rRNAs, can easily be generated locally on a personal computer. The software and supplementary data are available at http://genome.ucf.edu/PMFastR.