Genetic algorithm-based feature selection with manifold learning for cancer classification using microarray data.

BACKGROUND: Microarray data have been widely utilized for cancer classification. The main characteristic of microarray data is "large p and small n" in that data contain a small number of subjects but a large number of genes. It may affect the validity of the classification. Thus, there is a pressing demand of techniques able to select genes relevant to cancer classification. RESULTS: This study proposed a novel feature (gene) selection method, Iso-GA, for cancer classification. Iso-GA hybrids the manifold learning algorithm, Isomap, in the genetic algorithm (GA) to account for the latent nonlinear structure of the gene expression in the microarray data. The Davies-Bouldin index is adopted to evaluate the candidate solutions in Isomap and to avoid the classifier dependency problem. Additionally, a probability-based framework is introduced to reduce the possibility of genes being randomly selected by GA. The performance of Iso-GA was evaluated on eight benchmark microarray datasets of cancers. Iso-GA outperformed other benchmarking gene selection methods, leading to good classification accuracy with fewer critical genes selected. CONCLUSIONS: The proposed Iso-GA method can effectively select fewer but critical genes from microarray data to achieve competitive classification performance.

Nanopore basecalling from a perspective of instance segmentation.

BACKGROUND: Nanopore sequencing is a rapidly developing third-generation sequencing technology, which can generate long nucleotide reads of molecules within a portable device in real-time. Through detecting the change of ion currency signals during a DNA/RNA fragment's pass through a nanopore, genotypes are determined. Currently, the accuracy of nanopore basecalling has a higher error rate than the basecalling of short-read sequencing. Through utilizing deep neural networks, the-state-of-the art nanopore basecallers achieve basecalling accuracy in a range from 85% to 95%. RESULT: In this work, we proposed a novel basecalling approach from a perspective of instance segmentation. Different from previous approaches of doing typical sequence labeling, we formulated the basecalling problem as a multi-label segmentation task. Meanwhile, we proposed a refined U-net model which we call UR-net that can model sequential dependencies for a one-dimensional segmentation task. The experiment results show that the proposed basecaller URnano achieves competitive results on the in-species data, compared to the recently proposed CTC-featured basecallers. CONCLUSION: Our results show that formulating the basecalling problem as a one-dimensional segmentation task is a promising approach, which does basecalling and segmentation jointly.

Detection of APC mosaicism by next-generation sequencing in an FAP patient.

Familial adenomatous polyposis (FAP) of the colon is characterized by multiple polyps in the intestine and extra-colonic manifestations. Most FAP cases are caused by a germline mutation in the tumor-suppressor gene APC, but some cases of adenomatous polyposis result from germline mutations in MUTYH, POLD1 or POLE. Although sequence analysis of APC by the Sanger method is routinely performed for genetic testing, there remain cases whose mutations are not detected by the analysis. Next-generation sequencing has enabled us to analyze the comprehensive human genome, improving the chance of identifying disease causative variants. In this study, we conducted whole-genome sequencing of a sporadic FAP patient in which we did not find any pathogenic APC mutations by the conventional Sanger sequencing. Whole-genome sequencing and subsequent deep sequencing identified a mosaic mutation of c.3175G>T, p.E1059X in ~12% of his peripheral leukocytes. Additional deep sequencing of his buccal mucosa, hair follicles, non-cancerous mucosa of the stomach and colon disclosed that these tissues harbored the APC mutation at different frequencies. Our data implied that genetic analysis by next-generation sequencing is an effective strategy to identify genetic mosaicism in hereditary diseases.

Serum total bilirubin elevation is a predictor of the clinicopathological severity of acute appendicitis.

PURPOSE: Elevation of the serum total bilirubin (STB) level not stemming from hepatic dysfunction or biliary obstruction may be seen in cases of acute appendicitis. This paper deals with the clinical significance of such elevations. METHODS: Data from 410 appendectomized patients classified into two groups (a high preoperative STB group and a normal preoperative STB group) were analyzed to reveal the significance of preoperative hyperbilirubinemia. We also examined whether the preoperative STB level might serve as a risk factor for gangrenous appendicitis by a multivariate analysis. RESULTS: Gangrenous appendicitis was more common in the high preoperative STB group (p < 0.001). The multivariate analysis revealed that an elevated preoperative STB level (odds ratio 1.7919) was a risk factor for gangrenous appendicitis. CONCLUSION: In patients with an elevated preoperative STB level, it is very likely that the inflammation is severe and that the disease has progressed to a severe condition histopathologically; therefore, meticulous attention should be paid to the selection of the surgical procedure, as well as to the postoperative clinical course.

Electron microscopic structures, serum resistance, and plasmid restructuring of New Delhi metallo-ß-lactamase-1 (NDM-1)-producing ST42 Klebsiella pneumoniae emerging in Japan.

Enterobacteriaceae, carrying the New Delhi metallo-ß-lactamase-1 (NDM-1) gene (bla (NDM-1)), have emerged and posed a threat since 2006. In Japan, bla (NDM-1)-carrying Escherichia coli was first described in 2010. In this study, we characterized NDM-1-positive Klebsiella pneumoniae strain 419 in Japan, which was isolated from the urine of a 90-year-old Japanese patient who had never been to the Indian subcontinent. K. pneumoniae 419 belonged to ST42. It possessed a surface capsule (with untypeable capsular PCR types) and was resistant to serum killing. K. pneumoniae 419 cells were occasionally flagellated or piliated and autoaggregated. K. pneumoniae 419 was resistant to ß-lactams (including carbapenems), aminoglycosides, and fluoroquinolones, and was susceptible to imipenem (or biapenem), aztreonam, polymixin B, and colistin. It possessed at least eight plasmids; of those, a 74-kb plasmid (pKPJ1) of the replicon FIIA carried bla (NDM-1) and was conjugally transferred to E. coli strains, with a 71-kb transferable azithromycin-resistant (mphA (+)) plasmid of the replicon F (pKPJ2), as a large (145-kb) plasmid (pKPJF100) through a transposition event. In addition to bla (NDM-1), pKPJ1 carried arr-2, pKPJ2 carried mphA, and pKPJF100 carried both. They were negative for the 16S rRNA methylase gene, e.g., which is frequently associated with bla (NDM-1). The data demonstrate that K. pneumoniae 419 possessed virulence- and fitness-associated surface structures, was resistant to serum killing, and possessed a unique (or rare) genetic background in terms of ST type and bla (NDM-1)-carrying plasmid.

Privacy-Preserving Statistical Analysis of Genomic Data Using Compressive Mechanism with Haar Wavelet Transform.

To promote the use of personal genome information in medicine, it is important to analyze the relationship between diseases and the human genomes. Therefore, statistical analysis using genomic data is often conducted, but there is a privacy concern with respect to releasing the statistics as they are. Existing methods to address this problem using the concept of differential privacy cannot provide accurate outputs under strong privacy guarantees, making them less practical. In this study, for the first time, we investigate the application of a compressive mechanism to genomic statistical data and propose two approaches. The first is to apply the normal compressive mechanism to the statistics vector along with an algorithm to determine the number of nonzero entries in a sparse representation. The second is to alter the mechanism based on the data, aiming to release significant single nucleotide polymorphisms with a high probability. In this algorithm, we apply the compressive mechanism with the input as a sparse vector for significant data and the Laplace mechanism for nonsignificant data. By using the Haar wavelet transform for the compressive mechanism, we can determine the number of nonzero elements and the amount of noise. In addition, we give theoretical guarantees that our proposed methods achieve ϵ-differential privacy. We evaluated our methods in terms of accuracy and rank error compared with the Laplace and exponential mechanisms. The results show that our second method in particular can guarantee high privacy assurance as well as utility.

Efficient Differentially Private Methods for a Transmission Disequilibrium Test in Genome Wide Association Studies.

To achieve the provision of personalized medicine, it is very important to investigate the relationship between diseases and human genomes. For this purpose, large-scale genetic studies such as genome-wide association studies are often conducted, but there is a risk of identifying individuals if the statistics are released as they are. In this study, we propose new efficient differentially private methods for a transmission disequilibrium test, which is a family-based association test. Existing methods are computationally intensive and take a long time even for a small cohort. Moreover, for approximation methods, sensitivity of the obtained values is not guaranteed. We present an exact algorithm with a time complexity of 𝒪(nm) for a dataset containing n families and m single nucleotide polymorphisms (SNPs). We also propose an approximation algorithm that is faster than the exact one and prove that the obtained scores' sensitivity is 1. From our experimental results, we demonstrate that our exact algorithm is 10, 000 times faster than existing methods for a small cohort with 5, 000 SNPs. The results also indicate that the proposed method is the first in the world that can be applied to a large cohort, such as those with 106 SNPs. In addition, we examine a suitable dataset to apply our approximation algorithm. Supplementary materials are available at https://github.com/ay0408/DP-trio-TDT.

More practical differentially private publication of key statistics in GWAS.

Motivation: Analyses of datasets that contain personal genomic information are very important for revealing associations between diseases and genomes. Genome-wide association studies, which are large-scale genetic statistical analyses, often involve tests with contingency tables. However, if the statistics obtained by these tests are made public as they are, sensitive information of individuals could be leaked. Existing studies have proposed privacy-preserving methods for statistics in the χ2 test with a 3 × 2 contingency table, but they do not cover all the tests used in association studies. In addition, existing methods for releasing differentially private P-values are not practical. Results: In this work, we propose methods for releasing statistics in the χ2 test, the Fisher's exact test and the Cochran-Armitage's trend test while preserving both personal privacy and utility. Our methods for releasing P-values are the first to achieve practicality under the concept of differential privacy by considering their base 10 logarithms. We make theoretical guarantees by showing the sensitivity of the above statistics. From our experimental results, we evaluate the utility of the proposed methods and show appropriate thresholds with high accuracy for using the private statistics in actual tests. Availability and implementation: A python implementation of our experiments is available at https://github.com/ay0408/DP-statistics-GWAS. Supplementary information: Supplementary data are available at Bioinformatics Advances online.

New kernel methods for phenotype prediction from genotype data.

Phenotype prediction from genotype data is one of the most important issues in computational genetics. In this work, we propose a new kernel (i.e., an SVM: Support Vector Machine) method for phenotype prediction from genotype data. In our method, we first infer multiple suboptimal haplotype candidates from each genotype by using the HMM (Hidden Markov Model), and the kernel matrix is computed based on the predicted haplotype candidates and their emission probabilities from the HMM. We validated the performance of our method through experiments on several datasets: One is an artificially constructed dataset via a program GeneArtisan, others are a real dataset of the NAT2 gene from the international HapMap project, and a real dataset of genotypes of diseased individuals. The experiments show that our method is superior to ordinary naive kernel methods (i.e., not based on haplotype prediction), especially in cases of strong LD (linkage disequilibrium).

Two successful curative operations using stomach-preserving distal pancreatectomy with celiac axis resection for the treatment of locally advanced pancreatic body cancer.

Large vessel invasion is a serious factor determining whether an operation for pancreatic body cancer is feasible. The Appleby operation is a radical operation for the treatment of pancreatic body cancer that has infiltrated the celiac axis. Since this procedure includes a total gastrectomy, the operation is associated with a high morbidity, mortality, and deteriorating postoperative quality of life (QOL). We experienced two cases in which radical operations consisting of a stomach-preserving distal pancreatectomy with en bloc resection of the celiac, common hepatic, and left gastric artery were performed. The use of adjuvant chemotherapy in these cases led to a good postoperative QOL.

Efficient substructure RMSD query algorithms.

Protein structure analysis is a very important research topic in the molecular biology of the post-genomic era. The root mean square deviation (RMSD) is the most frequently used measure for comparing two protein three-dimensional (3-D) structures. In this paper, we deal with two fundamental problems related to the RMSD. We first deal with a problem called the "range RMSD query" problem. Given an aligned pair of structures, the problem is to compute the RMSD between two aligned substructures of them without gaps. This problem has many applications in protein structure analysis. We propose a linear-time preprocessing algorithm that enables constant-time RMSD computation. Next, we consider a problem called the "substructure RMSD query" problem, which is a generalization of the above range RMSD query problem. It is a problem to compute the RMSD between any substructures of two unaligned structures without gaps. Based on the algorithm for the range RMSD problem, we propose an O(nm) preprocessing algorithm that enables constant-time RMSD computation, where n and m are the lengths of the given structures. Moreover, we propose O(nm log r/r)-time and O(nm/r)-space preprocessing algorithm that enables O(r) query, where r is an arbitrary integer such that 1 < or = r < or = min(n, m). We also show that our strategy also works for another measure called the unit-vector root mean square deviation (URMSD), which is a variant of the RMSD.

Successful treatment of a colonic ulcer penetrating the urinary bladder caused by the administration of calcium polystyrene sulfonate and sorbitol.

A 77-year-old woman was urgently admitted for the treatment of diabetic ketoacidosis and a duodenal ulcer hemorrhage in March 1999. She had a history of diabetes and angina pectoris. After admission, she received oral calcium polystyrene sulfonate and sorbitol to treat hyperkalemia. Nine days later, severe abdominal pain developed. A colonoscopic examination revealed a sigmoid colonic ulcer and stenosis; the patient was treated conservatively. At a 1-year follow-up examination, the colonic stenosis was found have worsened; pneumaturia developed in January 2001. The patient was found to have a sigmoidovesical fistula and underwent sigmoidectomy and partial resection of the ileum and urinary bladder. The histological findings were a benign colonic ulcer with the infiltration of inflammatory cells, mainly lymphocytes. Rhomboidal, dark violet Kayexalate crystals were observed on microscope examination in the submucosa in both the first and second colonic biopsy specimens. We concluded that the colonic ulcer and the sigmoidovesical fistula had been caused by the administration of calcium polystyrene sulfonate and sorbitol. Reports of colonic perforation as a result of the administration of calcium polystyrene sulfonate and sorbitol are rare. Here, we report the successful treatment of a colonic ulcer that had penetrated the urinary bladder.

Significance of gastrectomy as palliative surgery for gastric carcinoma with pyloric stenosis.

We conducted a clinicopathological study of cases of gastric carcinoma with pyloric stenosis and examined treatment outcomes and the prognosis of cases of stage IV gastric carcinoma with pyloric stenosis and the validity of gastrectomy as palliative surgery in these cases. The outcomes of 49 surgeries for gastric carcinoma with pyloric stenosis were compared with those of 671 surgeries for gastric carcinoma without pyloric stenosis. The diagnosis of pyloric stenosis was confirmed with both upper gastrointestinal endoscopy and an upper gastrointestinal barium series. The frequency of pyloric stenosis in patients with gastric carcinoma was 7.3%. Serosal invasion was observed in about 70% of all cases. Of these cases, 53.1% were classified as stage IV. The resection rate was 73.5%, and the resection was classified as curative in 44.9% of cases. The incidence of complications after surgery in cases of stage IV gastric carcinoma was 47.1%. The median survival time was significantly greater in patients undergoing resection group than in those not undergoing resection (p=0.025). Most patients with gastric cancer and pyloric stenosis can be considered to have stage IV disease, which is associated with high rates of morbidity and mortality; thus, prevention of complications, and therefore, avoidance of gastrectomy is recommended in such patients. Nonetheless, in this study, gastrectomy was shown to improve prognoses in these patients.

Revealing phenotype-associated functional differences by genome-wide scan of ancient haplotype blocks.

Genome-wide scans for positive selection have become important for genomic medicine, and many studies aim to find genomic regions affected by positive selection that are associated with risk allele variations among populations. Most such studies are designed to detect recent positive selection. However, we hypothesize that ancient positive selection is also important for adaptation to pathogens, and has affected current immune-mediated common diseases. Based on this hypothesis, we developed a novel linkage disequilibrium-based pipeline, which aims to detect regions associated with ancient positive selection across populations from single nucleotide polymorphism (SNP) data. By applying this pipeline to the genotypes in the International HapMap project database, we show that genes in the detected regions are enriched in pathways related to the immune system and infectious diseases. The detected regions also contain SNPs reported to be associated with cancers and metabolic diseases, obesity-related traits, type 2 diabetes, and allergic sensitization. These SNPs were further mapped to biological pathways to determine the associations between phenotypes and molecular functions. Assessments of candidate regions to identify functions associated with variations in incidence rates of these diseases are needed in the future.

Dictionary-driven prokaryotic gene finding.

Gene identification, also known as gene finding or gene recognition, is among the important problems of molecular biology that have been receiving increasing attention with the advent of large scale sequencing projects. Previous strategies for solving this problem can be categorized into essentially two schools of thought: one school employs sequence composition statistics, whereas the other relies on database similarity searches. In this paper, we propose a new gene identification scheme that combines the best characteristics from each of these two schools. In particular, our method determines gene candidates among the ORFs that can be identified in a given DNA strand through the use of the Bio-Dictionary, a database of patterns that covers essentially all of the currently available sample of the natural protein sequence space. Our approach relies entirely on the use of redundant patterns as the agents on which the presence or absence of genes is predicated and does not employ any additional evidence, e.g. ribosome-binding site signals. The Bio-Dictionary Gene Finder (BDGF), the algorithm's implementation, is a single computational engine able to handle the gene identification task across distinct archaeal and bacterial genomes. The engine exhibits performance that is characterized by simultaneous very high values of sensitivity and specificity, and a high percentage of correctly predicted start sites. Using a collection of patterns derived from an old (June 2000) release of the Swiss-Prot/TrEMBL database that contained 451 602 proteins and fragments, we demonstrate our method's generality and capabilities through an extensive analysis of 17 complete archaeal and bacterial genomes. Examples of previously unreported genes are also shown and discussed in detail.

The web server of IBM's Bioinformatics and Pattern Discovery group.

We herein present and discuss the services and content which are available on the web server of IBM's Bioinformatics and Pattern Discovery group. The server is operational around the clock and provides access to a variety of methods that have been published by the group's members and collaborators. The available tools correspond to applications ranging from the discovery of patterns in streams of events and the computation of multiple sequence alignments, to the discovery of genes in nucleic acid sequences and the interactive annotation of amino acid sequences. Additionally, annotations for more than 70 archaeal, bacterial, eukaryotic and viral genomes are available on-line and can be searched interactively. The tools and code bundles can be accessed beginning at http://cbcsrv.watson.ibm.com/Tspd.html whereas the genomics annotations are available at http://cbcsrv.watson.ibm.com/Annotations/.

Reduced expression of APC-1B but not APC-1A by the deletion of promoter 1B is responsible for familial adenomatous polyposis.

Germline mutations in the tumor suppressor gene APC are associated with familial adenomatous polyposis (FAP). Here we applied whole-genome sequencing (WGS) to the DNA of a sporadic FAP patient in which we did not find any pathological APC mutations by direct sequencing. WGS identified a promoter deletion of approximately 10 kb encompassing promoter 1B and exon1B of APC. Additional allele-specific expression analysis by deep cDNA sequencing revealed that the deletion reduced the expression of the mutated APC allele to as low as 11.2% in the total APC transcripts, suggesting that the residual mutant transcripts were driven by other promoter(s). Furthermore, cap analysis of gene expression (CAGE) demonstrated that the deleted promoter 1B region is responsible for the great majority of APC transcription in many tissues except the brain. The deletion decreased the transcripts of APC-1B to 39-45% in the patient compared to the healthy controls, but it did not decrease those of APC-1A. Different deletions including promoter 1B have been reported in FAP patients. Taken together, our results strengthen the evidence that analysis of structural variations in promoter 1B should be considered for the FAP patients whose pathological mutations are not identified by conventional direct sequencing.

Whole-genome mutational landscape and characterization of noncoding and structural mutations in liver cancer.

Liver cancer, which is most often associated with virus infection, is prevalent worldwide, and its underlying etiology and genomic structure are heterogeneous. Here we provide a whole-genome landscape of somatic alterations in 300 liver cancers from Japanese individuals. Our comprehensive analysis identified point mutations, structural variations (STVs), and virus integrations, in noncoding and coding regions. We discovered mutational signatures related to liver carcinogenesis and recurrently mutated coding and noncoding regions, such as long intergenic noncoding RNA genes (NEAT1 and MALAT1), promoters, CTCF-binding sites, and regulatory regions. STV analysis found a significant association with replication timing and identified known (CDKN2A, CCND1, APC, and TERT) and new (ASH1L, NCOR1, and MACROD2) cancer-related genes that were recurrently affected by STVs, leading to altered expression. These results emphasize the value of whole-genome sequencing analysis in discovering cancer driver mutations and understanding comprehensive molecular profiles of liver cancer, especially with regard to STVs and noncoding mutations.

Keratinocyte growth factor receptor expression in normal colorectal epithelial cells and differentiated type of colorectal cancer.

The keratinocyte growth factor receptor, also known as KGFR/FGFR2 IIIb, is mainly localized in epithelial cells and participates in the proliferation of these cells. In the present study, we attempted to clarify the expression and role of KGFR in human colorectal cancer. The KGFR protein was detected in several colorectal cancer cell lines. Immuno-histochemically, KGFR was expressed on the luminal surface of epithelial cells in normal colorectal tissues. KGFR was expressed in 35 of 56 (62.5%) colorectal cancer patients and localized at the center of the cancer nests. Cytokeratin 20 (CK 20) colocalized with KGFR. In contrast, Ki-67 localized at the periphery of the cancer nests. Clinicopathologically, a high level of KGFR expression was associated with well-differentiated histological type (p<0.0001) and shallow wall invasion (p<0.0174). These findings indicate that KGFR may play important roles in the differentiation of normal colorectal epithelial cells and establishment of the well-differentiated histological type of colorectal cancer cells.

An O(m log m)-Time Algorithm for Detecting Superbubbles.

In genome assembly graphs, motifs such as tips, bubbles, and cross links are studied in order to find sequencing errors and to understand the nature of the genome. Superbubble, a complex generalization of bubbles, was recently proposed as an important subgraph class for analyzing assembly graphs. At present, a quadratic time algorithm is known. This paper gives an O(m log m)-time algorithm to solve this problem for a graph with m edges.