This preliminary study shows that array-CGH is useful for detecting CNVs in cases of RPL. Further investigations of CNVs, particularly those involving genes that are imprinted in placenta, in women with RPL could be worthwhile.

Developers of high-performance algorithms for hard computational problems increasingly take advantage of automated algorithm configuration tools, and consequently often create solvers with many parameters and vast configuration spaces. However, there has been very little work to help these algorithm developers answer questions about the high-quality configurations produced by these tools, specifically about which parameter changes contribute most to improved performance. In this work, we present an automated technique for answering such questions by performing ablation analysis between two algorithm configurations. We perform an extensive empirical analysis of our technique on five scenarios from propositional satisfiability, mixed-integer programming and AI planning, and show that in all of these scenarios more than 95% of the performance gains between default configurations and configurations obtained by using automated configuration tools can be explained by modifying the values of a small number of parameters (1-4 in the scenarios we studied).

No abstract

Higher resolution whole-genome arrays facilitate the identification of smaller copy number variations (CNVs) and their integral genes contributing to autism and/or intellectual disability (ASD/ID). Our study describes the use of one of the highest resolution arrays, the Affymetrix(®) Cytogenetics 2.7M array, coupled with quantitative multiplex polymerase chain reaction (PCR) of short fluorescent fragments (QMPSF) for detection and validation of small CNVs. We studied 82 subjects with ASD and ID in total (30 in the validation and 52 in the application cohort) and detected putatively pathogenic CNVs in 6/52 cases from the application cohort. This included a 130-kb maternal duplication spanning exons 64-79 of the DMD gene which was found in a 3-year-old boy manifesting autism and mild neuromotor delays. Other pathogenic CNVs involved 4p14, 12q24.31, 14q32.31, 15q13.2-13.3, and 17p13.3. We established the optimal experimental conditions which, when applied to select small CNVs for QMPSF confirmation, reduced the false positive rate from 60% to 25%. Our work suggests that selection of small CNVs based on the function of integral genes, followed by review of array experimental parameters resulting in highest confirmation rate using multiplex PCR, may enhance the usefulness of higher resolution platforms for ASD and ID gene discovery.

Developmental abnormalities of human embryos can be visualized in utero using embryoscopy. Our previous embryoscopic and genetic evaluations detected developmental abnormalities in the majority of both euploid (74%) and aneuploid or polyploid (90%) miscarriages. Since we found the pattern of morphological changes to be similar in euploid and non-euploid embryos, we proposed that lethal submicroscopic changes, not detected by standard chromosome testing, may be responsible for miscarriage of euploid embryos. Whole genome oligo and bacterial artificial chromosome array comparative genome hybridization (CGH) was used to screen for submicroscopic chromosomal changes (DNA copy number variants or CNVs) in 17 euploid embryonic miscarriages, with a range of developmental abnormalities documented by embryoscopy. The CNV breakpoints were refined using a custom array (Agilent) with high resolution coverage of the CNVs. Six unique CNVs, previously not reported, were identified in 5 of the 17 embryos (29% of all cases or 50% of cases studied with higher resolution arrays). All six unique CNVs were <250 kb in size. On the basis of parental array CGH analysis, a de novo origin of a CNV was determined for one embryo (at 13q32.1) and suspected for another (at 10p15.3). Three CNVs, at Xq28, 1q25.3 and 7p14.3, were inherited and a CNV at 17p13.1 was of unknown origin. The genes contained within these unique CNVs will be discussed, with specific reference to rearrangements of syntaxin and tryptophan-aspartic acid (WD) repeat genes. Our report describes for the first time, de novo and inherited unique CNVs in euploid human embryos with specific developmental defects.

State-of-the-art planners often exhibit substantial runtime variation, making it useful to be able to efficiently predict how long a given planner will take to run on a given instance. In other areas of AI, such needs are met by building so-called empirical performance models (EPMs), statistical models derived from sets of problem instances and performance observations. Historically, such models have been less accurate for predicting the running times of planners. A key hurdle has been a relative weakness in instance features for characterizing instances: mappings from problem instances to real numbers that serve as the starting point for learning an EPM. We propose a new, extensive set of instance features for planning, and investigate its effectiveness across a range of model families. We built EPMs for various prominent planning systems on several thousand benchmark problems from the planning literature and from IPC benchmark sets, and conclude that our models predict runtime much more accurately than the previous state of the art. We also study the relative importance of these features.

Array CGH enables the detection of pathogenic copy number variants (CNVs) in 5-15% of individuals with intellectual disability (ID), making it a promising tool for uncovering ID candidate genes. However, most CNVs encompass multiple genes, making it difficult to identify key disease gene(s) underlying ID etiology. Using array CGH we identified 47 previously unreported unique CNVs in 45/255 probands. We prioritized ID candidate genes using five bioinformatic gene prioritization web tools. Gene priority lists were created by comparing integral genes from each CNV from our ID cohort with sets of training genes specific either to ID or randomly selected. Our findings suggest that different training sets alter gene prioritization only moderately; however, only the ID gene training set resulted in significant enrichment of genes with nervous system function (19%) in prioritized versus non-prioritized genes from the same de novo CNVs (7%, p < 0.05). This enrichment further increased to 31% when the five web tools were used in concert and included genes within mitogen-activated protein kinase (MAPK) and neuroactive ligand-receptor interaction pathways. Gene prioritization web tools enrich for genes with relevant function in ID and more readily facilitate the selection of ID candidate genes for functional studies, particularly for large CNVs.

Abstract. Sophisticated empirical methods drive the development of high-performance solvers for an increasing range of problems from industry and academia. However, automated tools implementing these methods are often difficult to develop and to use. We address this issue with two contributions. First, we develop a formal description of meta-algorithmic problems and use it as the basis for an automated algorithm analysis and design framework called the High-performance Algorithm Laboratory. Second, we describe HAL 1.0, an implementation of the core components of this framework that provides support for distributed execution, remote monitoring, data management, and analysis of results. We demonstrate our approach by using HAL 1.0 to conduct a sequence of increasingly complex analysis and design tasks on state-of-the-art solvers for SAT and mixed-integer programming problems.