Machine Learning and Deep Learning Approaches in Breast Cancer Survival Prediction Using Clinical Data.

Breast cancer survival prediction can have an extreme effect on selection of best treatment protocols. Many approaches such as statistical or machine learning models have been employed to predict the survival prospects of patients, but newer algorithms such as deep learning can be tested with the aim of improving the models and prediction accuracy. In this study, we used machine learning and deep learning approaches to predict breast cancer survival in 4,902 patient records from the University of Malaya Medical Centre Breast Cancer Registry. The results indicated that the multilayer perceptron (MLP), random forest (RF) and decision tree (DT) classifiers could predict survivorship, respectively, with 88.2 %, 83.3 % and 82.5 % accuracy in the tested samples. Support vector machine (SVM) came out to be lower with 80.5 %. In this study, tumour size turned out to be the most important feature for breast cancer survivability prediction. Both deep learning and machine learning methods produce desirable prediction accuracy, but other factors such as parameter configurations and data transformations affect the accuracy of the predictive model.

How automated image analysis techniques help scientists in species identification and classification?

Identification of taxonomy at a specific level is time consuming and reliant upon expert ecologists. Hence the demand for automated species identification incre-ased over the last two decades. Automation of data classification is primarily focussed on images while incorporating and analysing image data has recently become easier due to developments in computational technology. Research ef-forts on identification of species include specimens' image processing, extraction of identical features, followed by classifying them into correct categories. In this paper, we discuss recent automated species identification systems, mainly for categorising and evaluating their methods. We reviewed and compared different methods in step by step scheme of automated identification and classification systems of species images. The selection of methods is influenced by many variables such as level of classification, number of training data and complexity of images. The aim of writing this paper is to provide researchers and scientists an extensive background study on work related to automated species identification, focusing on pattern recognition techniques in building such systems for biodiversity studies. (Folia Morphol 2018; 77, 2: 179-193).

Transcriptome sequencing to produce SNP-based genetic maps of onion.

We used the Roche-454 platform to sequence from normalized cDNA libraries from each of two inbred lines of onion (OH1 and 5225). From approximately 1.6 million reads from each inbred, 27,065 and 33,254 cDNA contigs were assembled from OH1 and 5225, respectively. In total, 3,364 well supported single nucleotide polymorphisms (SNPs) on 1,716 cDNA contigs were identified between these two inbreds. One SNP on each of 1,256 contigs was randomly selected for genotyping. OH1 and 5225 were crossed and 182 gynogenic haploids extracted from hybrid plants were used for SNP mapping. A total of 597 SNPs segregated in the OH1 × 5225 haploid family and a genetic map of ten linkage groups (LOD ≥8) was constructed. Three hundred and thirty-nine of the newly identified SNPs were also mapped using a previously developed segregating family from BYG15-23 × AC43, and 223 common SNPs were used to join the two maps. Because these new SNPs are in expressed regions of the genome and commonly occur among onion germplasms, they will be useful for genetic mapping, gene tagging, marker-aided selection, quality control of seed lots, and fingerprinting of cultivars.

Functional analysis of a TGA factor-binding site located in the promoter region controlling salicylic acid-induced NIMIN-1 expression in Arabidopsis.

TGA factors play a key role in plant defense by binding to the promoter region of defense genes, inducing expression. Salicylic acid (SA) induces the expression of the gene encoding NIMIN-1, which interacts with NPR1/NIM1, a key regulator of systemic acquired resistance. We investigated whether the TGA2-binding motif TGACG located upstream of the NIMIN-1 gene is necessary for SA induction of NIMIN-1 expression. A mutated version of the NIMIN-1 promoter was created by site-directed mutagenesis. We generated T-DNA constructs in which native NIMIN-1 and mutated promoters were fused to green fluorescent protein and beta-glucuronidase reporters. We produced transgenic Arabidopsis plants and observed NIMIN-1 promoter-driven green fluorescent protein expression in the roots, petiole and leaves. Constructs were agroinfiltrated into the leaves for transient quantitative assays of gene expression. Using quantitative real-time RT-PCR, we characterized the normal gene response to SA and compared it to the response of the mutant version of the NIMIN-1 promoter. Both the native NIMIN-1 construct and an endogenous copy of NIMIN-1 were induced by SA. However, the mutated promoter construct was much less sensitive to SA than the native NIMIN-1 promoter, indicating that this TGA2-binding motif is directly involved in the modulation of SA-induced NIMIN-1 expression in Arabidopsis.

Intracellular pH in Dictyostelium discoideum: a 31P nuclear magnetic resonance study.

We have used phosphorus-31 nuclear magnetic resonance to determine intracellular pH in the cellular slime mold Dictyostelium discoideum. We devised an air-lift circulator to maintain the dense cell suspensions in a well-oxygenated and well-stirred state while causing minimal perturbation to the sample flowing through the detector coils. Cells continued to develop normally in this set-up. Spectra acquired under these conditions typically show two peaks in the inorganic phosphate region corresponding to pH values of 7.16 +/- 0.03 and 6.48 +/- 0.02. These peaks are believed to represent the mitochondrial and cytosolic compartments respectively, based on a comparison of these values with published data and the collapse of the two compartments upon addition of the mitochondrial uncoupler carbonyl cyanide 4-(trifluoromethoxy)-phenylhydrazone. Dictyostelium cells show a remarkable degree of intracellular pH homeostasis. Both mitochondrial and cytosolic pH remained unchanged as extracellular pH was varied from 4.3 to 8.1. There was also no apparent change in the pH of either compartment after up to 13.5 hours' development in suspension.

DNA polymerase required for rapid repair of x-ray--induced DNA strand breaks in vivo.

A much higher yield of DNA single-strand breaks was obtained in the DNA polymerase-deficient mutant Escherichia coli K-12 pol A1 after a given dose of x-rays than had been found before in Escherichia coli. The increased yield of single-strand breaks was due to the absence of a rapid repair system, which had not been described in Escherichia coli K-12. This absence probably accounts for the x-ray sensitivity of the pol A1 mutant. The rapid repair system can be reversibly inhibited in pol+ cells.

Isolation and characterization of the highly repeated fraction of the banana genome.

Although the nuclear genome of banana (Musa spp.) is relatively small (1C approximately 610 Mbp for M. acuminata), the results obtained from other sequenced genomes suggest that more than half of the banana genome may be composed of repetitive and non-coding DNA sequences. Knowledge of repetitive DNA can facilitate mapping of important traits, phylogenetic studies, BAC-based physical mapping, and genome sequencing/annotation. However, only a few repetitive DNA sequences have been characterized in banana. In this work, we used DNA reassociation kinetics to isolate the highly repeated fraction of the banana genome (M. acuminata 'Calcutta 4'). Two libraries, one prepared from Cot

The Arabidopsis Information Resource (TAIR): a comprehensive database and web-based information retrieval, analysis, and visualization system for a model plant.

Arabidopsis thaliana, a small annual plant belonging to the mustard family, is the subject of study by an estimated 7000 researchers around the world. In addition to the large body of genetic, physiological and biochemical data gathered for this plant, it will be the first higher plant genome to be completely sequenced, with completion expected at the end of the year 2000. The sequencing effort has been coordinated by an international collaboration, the Arabidopsis Genome Initiative (AGI). The rationale for intensive investigation of Arabidopsis is that it is an excellent model for higher plants. In order to maximize use of the knowledge gained about this plant, there is a need for a comprehensive database and information retrieval and analysis system that will provide user-friendly access to Arabidopsis information. This paper describes the initial steps we have taken toward realizing these goals in a project called The Arabidopsis Information Resource (TAIR) (www.arabidopsis.org).

Sequence and organization of 5S ribosomal RNA-encoding genes of Arabidopsis thaliana.

We have isolated a genomic clone containing Arabidopsis thaliana 5S ribosomal RNA (rRNA)-encoding genes (rDNA) by screening an A. thaliana library with a 5S rDNA probe from flax. The clone isolated contains seven repeat units of 497 bp, plus 11 kb of flanking genomic sequence at one border. Sequencing of individual subcloned repeat units shows that the sequence of the 5S rRNA coding region is very similar to that reported for other flowering plants. Four A. thaliana ecotypes were found to contain approx. 1000 copies of 5S rDNA per haploid genome. Southern-blot analysis of genomic DNA indicates that 5S rDNA occurs in long tandem arrays, and shows the presence of numerous restriction-site polymorphisms among the six ecotypes studied.

Use of RNA and genomic DNA references for inferred comparisons in DNA microarray analyses.

In most microarray assays, labeled cDNA molecules derived from reference and query RNA samples are co-hybridized to probes arrayed on a glass surface. Gene expression profiles are then calculated for each gene based on the relative hybridization intensities measured between the two samples. The most commonly used reference samples are typically isolates from a single representative RNA source (RNA-0) or pooled mixtures of RNA derived from a plurality of sources (RNA-p). Genomic DNA offers an alternative reference nucleic acid with a number of potential advantages, including stability, reproducibility, and a potentially uniform representation of all genes, as each unique gene should have equal representation in a haploid genome. Using hydrogen peroxide-treated Arabidopsis thaliana plants as a model, we evaluated genomic DNA and RNA-p as reference samples and compared expression levels inferred through the reference relative to unexposed plants with expression levels measured directly using an RNA-0 reference. Our analysis demonstrates that while genomic DNA can serve as a reasonable reference source for microarray assays, a much greater correlation with direct measurements can be achieved using an RNA-based reference sample.

Deletion analysis of the essentiality of penicillin-binding proteins 1A, 2B and 2X of Streptococcus pneumoniae.

An internal fragment from each of the penicillin-binding protein (PBP) 1A, 2B and 2X genes of Streptococcus pneumoniae, which included the region encoding the active-site serine residue, was replaced by a fragment encoding spectinomycin resistance. The resulting constructs were tested for their ability to transform S. pneumoniae strain R6 to spectinomycin resistance. Spectinomycin-resistant transformants could not be obtained using either the inactivated PBP 2X or 2B genes, suggesting that deletion of either of these genes was a lethal event, but they were readily obtained using the inactivated PBP 1A gene. Analysis using the polymerase chain reaction confirmed that the latter transformants had replaced their chromosomal copy of the PBP 1A gene with the inactivated copy of the gene. Deletion of the PBP 1A gene was therefore tolerated under laboratory conditions and appeared to have little effect on growth or susceptibility to benzylpenicillin.

Pigeonpea genomics initiative (PGI): an international effort to improve crop productivity of pigeonpea (Cajanus cajan L.).

Pigeonpea (Cajanus cajan), an important food legume crop in the semi-arid regions of the world and the second most important pulse crop in India, has an average crop productivity of 780 kg/ha. The relatively low crop yields may be attributed to non-availability of improved cultivars, poor crop husbandry and exposure to a number of biotic and abiotic stresses in pigeonpea growing regions. Narrow genetic diversity in cultivated germplasm has further hampered the effective utilization of conventional breeding as well as development and utilization of genomic tools, resulting in pigeonpea being often referred to as an 'orphan crop legume'. To enable genomics-assisted breeding in this crop, the pigeonpea genomics initiative (PGI) was initiated in late 2006 with funding from Indian Council of Agricultural Research under the umbrella of Indo-US agricultural knowledge initiative, which was further expanded with financial support from the US National Science Foundation's Plant Genome Research Program and the Generation Challenge Program. As a result of the PGI, the last 3 years have witnessed significant progress in development of both genetic as well as genomic resources in this crop through effective collaborations and coordination of genomics activities across several institutes and countries. For instance, 25 mapping populations segregating for a number of biotic and abiotic stresses have been developed or are under development. An 11X-genome coverage bacterial artificial chromosome (BAC) library comprising of 69,120 clones have been developed of which 50,000 clones were end sequenced to generate 87,590 BAC-end sequences (BESs). About 10,000 expressed sequence tags (ESTs) from Sanger sequencing and ca. 2 million short ESTs by 454/FLX sequencing have been generated. A variety of molecular markers have been developed from BESs, microsatellite or simple sequence repeat (SSR)-enriched libraries and mining of ESTs and genomic amplicon sequencing. Of about 21,000 SSRs identified, 6,698 SSRs are under analysis along with 670 orthologous genes using a GoldenGate SNP (single nucleotide polymorphism) genotyping platform, with large scale SNP discovery using Solexa, a next generation sequencing technology, is in progress. Similarly a diversity array technology array comprising of ca. 15,000 features has been developed. In addition, >600 unique nucleotide binding site (NBS) domain containing members of the NBS-leucine rich repeat disease resistance homologs were cloned in pigeonpea; 960 BACs containing these sequences were identified by filter hybridization, BES physical maps developed using high information content fingerprinting. To enrich the genomic resources further, sequenced soybean genome is being analyzed to establish the anchor points between pigeonpea and soybean genomes. In addition, Solexa sequencing is being used to explore the feasibility of generating whole genome sequence. In summary, the collaborative efforts of several research groups under the umbrella of PGI are making significant progress in improving molecular tools in pigeonpea and should significantly benefit pigeonpea genetics and breeding. As these efforts come to fruition, and expanded (depending on funding), pigeonpea would move from an 'orphan legume crop' to one where genomics-assisted breeding approaches for a sustainable crop improvement are routine.

Sexual reproduction in the soybean sudden death syndrome pathogen Fusarium tucumaniae.

We investigated the sexual reproductive mode of the two most important etiological agents of soybean sudden death syndrome, Fusarium tucumaniae and Fusarium virguliforme. F. tucumaniae sexual crosses often were highly fertile, making it possible to assign mating type and assess female fertility in 24 South American isolates. These crosses produced red perithecia and oblong-elliptical ascospores, as is typical for sexual members of the F. solani species complex. Genotyping of progeny from three F. tucumaniae crosses confirmed that sexual recombination had occurred. In contrast, pairings among 17 U.S. F. virguliforme isolates never produced perithecia. Inter-species crosses between F. tucumaniae and F. virguliforme, in which infertile perithecia were induced only in one of the two F. tucumaniae mating types, suggest that all U.S. F. virguliforme isolates are of a single mating type. We conclude that the F. tucumaniae life cycle in S. America includes a sexual reproductive mode, and thus this species has greater potential for rapid evolution than the F. virguliforme population in the U.S., which may be exclusively asexual.

Genomic analysis of the rhg1 locus: candidate genes that underlie soybean resistance to the cyst nematode.

The rhg1 gene or genes lie at a recessive or co-dominant locus, necessary for resistance to all Hg types of the soybean (Glycine max (L.) Merr.) cyst nematode (Heterodera glycines I.). The aim here was to identify nucleotide changes within a candidate gene found at the rhg1 locus that were capable of altering resistance to Hg types 0 (race 3). A 1.5 +/- 0.25 cM region of chromosome 18 (linkage group G) was shown to encompass rhg1 using recombination events from four near isogenic line populations and nine DNA markers. The DNA markers anchored two bacterial artificial chromosome (BAC) clones 21d9 and 73p6. A single receptor like kinase (RLK; leucine rich repeat-transmembrane-protein kinase) candidate resistance gene was amplified from both BACs using redundant primers. The DNA sequence showed nine alleles of the RLK at Rhg1 in the soybean germplasm. Markers designed to detect alleles showed perfect association between allele 1 and resistance to soybean cyst nematode Hg types 0 in three segregating populations, fifteen additional selected recombination events and twenty-two Plant Introductions. A quantitative trait nucleotide (QTN) [corrected] in the RLK at rhg1 was inferred that alters A87 to V87 in the context of H274 rather than N274. [corrected] Contiguous DNA sequence of 315 kbp of chromosome 18 (about 2 cM) contained additional gene candidates that may modulate resistance to other Hg-types including a variant laccase, a hydrogen-sodium ion antiport and two proteins of unknown function. A molecular basis for recessive and co-dominant resistance that involves interactions among paralagous disease-resistance genes was inferred that would improve methods for developing new nematode-resistant soybean cultivars.

Genomic analysis of a region encompassing QRfs1 and QRfs2: genes that underlie soybean resistance to sudden death syndrome.

Candidate genes were identified for two loci, QRfs2 providing resistance to the leaf scorch called soybean (Glycine max (L.) Merr.) sudden death syndrome (SDS) and QRfs1 providing resistance to root infection by the causal pathogen Fusarium solani f.sp. glycines. The 7.5 +/- 0.5 cM region of chromosome 18 (linkage group G) was shown to encompass a cluster of resistance loci using recombination events from 4 near-isogenic line populations and 9 DNA markers. The DNA markers anchored 9 physical map contigs (7 are shown on the soybean Gbrowse, 2 are unpublished), 45 BAC end sequences (41 in Gbrowse), and contiguous DNA sequences of 315, 127, and 110 kbp. Gene density was high at 1 gene per 7 kbp only around the already sequenced regions. Three to 4 gene-rich islands were inferred to be distributed across the entire 7.5 cM or 3.5 Mbp showing that genes are clustered in the soybean genome. Candidate resistance genes were identified and a molecular basis for interactions among the disease resistance genes in the cluster inferred.