Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

The 1000 Genomes Project set out to provide a comprehensive description of common human genetic variation by applying whole-genome sequencing to a diverse set of individuals from multiple populations. Here we report completion of the project, having reconstructed the genomes of 2,504 individuals from 26 populations using a combination of low-coverage whole-genome sequencing, deep exome sequencing, and dense microarray genotyping. We characterized a broad spectrum of genetic variation, in total over 88 million variants (84.7 million single nucleotide polymorphisms (SNPs), 3.6 million short insertions/deletions (indels), and 60,000 structural variants), all phased onto high-quality haplotypes. This resource includes >99% of SNP variants with a frequency of >1% for a variety of ancestries. We describe the distribution of genetic variation across the global sample, and discuss the implications for common disease studies.

https://cdr.lib.unc.edu/downloads/v692t8920

A global reference for human genetic variation.

Comprehensive Integration of Single-Cell Data.

Genome-wide association studies have identified hundreds of genetic variants associated with complex human diseases and traits, and have provided valuable insights into their genetic architecture. Most variants identified so far confer relatively small increments in risk, and explain only a small proportion of familial clustering, leading many to question how the remaining, 'missing' heritability can be explained. Here we examine potential sources of missing heritability and propose research strategies, including and extending beyond current genome-wide association approaches, to illuminate the genetics of complex diseases and enhance its potential to enable effective disease prevention or treatment.

/pdf/finding-the-missing-heritability-of-complex-diseases-3b6mbqztnq.pdf

Finding the missing heritability of complex diseases

By characterizing the geographic and functional spectrum of human genetic variation, the 1000 Genomes Project aims to build a resource to help to understand the genetic contribution to disease. Here we describe the genomes of 1,092 individuals from 14 populations, constructed using a combination of low-coverage whole-genome and exome sequencing. By developing methods to integrate information across several algorithms and diverse data sources, we provide a validated haplotype map of 38 million single nucleotide polymorphisms, 1.4 million short insertions and deletions, and more than 14,000 larger deletions. We show that individuals from different populations carry different profiles of rare and common variants, and that low-frequency variants show substantial geographic differentiation, which is further increased by the action of purifying selection. We show that evolutionary conservation and coding consequence are key determinants of the strength of purifying selection, that rare-variant load varies substantially across biological pathways, and that each individual contains hundreds of rare non-coding variants at conserved sites, such as motif-disrupting changes in transcription-factor-binding sites. This resource, which captures up to 98% of accessible single nucleotide polymorphisms at a frequency of 1% in related populations, enables analysis of common and low-frequency variants in individuals from diverse, including admixed, populations.

/pdf/an-integrated-map-of-genetic-variation-from-1-092-human-4vxat9g1wd.pdf

An integrated map of genetic variation from 1,092 human genomes

Advances in high-throughput genotyping and the International HapMap Project have enabled association studies at the whole-genome level. We have constructed whole-genome genotyping panels of over 550,000 (HumanHap550) and 650,000 (HumanHap650Y) SNP loci by choosing tag SNPs from all populations genotyped by the International HapMap Project. These panels also contain additional SNP content in regions that have historically been overrepresented in diseases, such as nonsynonymous sites, the MHC region, copy number variant regions and mitochondrial DNA. We estimate that the tag SNP loci in these panels cover the majority of all common variation in the genome as measured by coverage of both all common HapMap SNPs and an independent set of SNPs derived from complete resequencing of genes obtained from SeattleSNPs. We also estimate that, given a sample size of 1,000 cases and 1,000 controls, these panels have the power to detect single disease loci of moderate risk (λ ∼ 1.8–2.0). Relative risks as low as λ ∼ 1.1–1.3 can be detected using 10,000 cases and 10,000 controls depending on the sample population and disease model. If multiple loci are involved, the power increases significantly to detect at least one locus such that relative risks 20%–35% lower can be detected with 80% power if between two and four independent loci are involved. Although our SNP selection was based on HapMap data, which is a subset of all common SNPs, these panels effectively capture the majority of all common variation and provide high power to detect risk alleles that are not represented in the HapMap data.

/pdf/power-to-detect-risk-alleles-using-genome-wide-tag-snp-4oqxro3lc6.pdf

Power to detect risk alleles using genome-wide tag SNP panels.

The present invention relates to methods for generating an array of amplified nucleic acid sequences. The methods can utilize amplicons that form nucleic acid balls that can be arrayed on a solid support. The invention additionally provides methods for obtaining targeted nucleic acid sequences.

Methods for generating amplified nucleic acid arrays

In this review, we describe the laboratory implementation of Infinium whole genome genotyping (WGG) technology for whole genome association studies and copy number studies. Briefly, the Infinium WGG assay employs a single tube whole genome amplification reaction to amplify the entire genome; genomic loci of interest are captured on an array by specific hybridization of picomolar concentrations amplified gDNA. After target capture, single nucleotide polymorphisms (SNPs) are genotyped on the array by a primer extension reaction using hapten-labeled nucleotides. The resultant hapten signal is amplified by immunhistochemical sandwich staining and the array is read out on a high resolution confocal scanner. We have combined this Infinium assay with high-density BeadChips to create the first array platform capable of genotyping over 1 million SNPs per slide. Additionally, the complete Infinium assay is automated using Tecan GenePaint slide processing system. Hybridization, washing, array-based primer extension and staining are performed directly in the Tecan capillary gap Te-Flow Through chambers. This automation process greatly increases assay robustness and throughput while enabling Laboratory Information Management System (LIMS) control of sample tracking. Finally, we give several examples of how this advance in genotyping technology is being applied in whole genome association and copy number studies.

Whole-genome genotyping on bead arrays.

To identify genes that are regulated by cis-acting functional elements in acute lymphoblastic leukemia (ALL) we determined the allele-specific expression (ASE) levels of 2529 genes by genotyping a genome-wide panel of single nucleotide polymorphisms in RNA and DNA from bone marrow and blood samples of 197 children with ALL. Using a reproducible, quantitative genotyping method and stringent criteria for scoring ASE, we found that 16% of the analyzed genes display ASE in multiple ALL cell samples. For most of the genes, the level of ASE varied largely between the samples, from 1.4-fold overexpression of one allele to apparent monoallelic expression. For genes exhibiting ASE, 55% displayed bidirectional ASE in which overexpression of either of the two SNP alleles occurred. For bidirectional ASE we also observed overall higher levels of ASE and correlation with the methylation level of these sites. Our results demonstrate that CpG site methylation is one of the factors that regulates gene expression in ALL cells.

/pdf/allele-specific-gene-expression-patterns-in-primary-leukemic-3fw7exnuyx.pdf

Allele-specific gene expression patterns in primary leukemic cells reveal regulation of gene expression by CpG site methylation

We report on a young male with moderate mental retardation, dysmorphic features, and language delay who is deleted for 7q31.1-7q31.31. His full karyotype is 46,XY,der(7)del(7)(q31.1q31.31)ins(10;7)(q24.3;q31.1q31.31)mat. This child had language impairment, including developmental verbal dyspraxia, but did not meet criteria for autism according to standardized ADOS testing. Our patient's deletion, which is the smallest reported deletion including FOXP2, adds to the body of evidence that supports the role of FOXP2 in speech and language impairment, but not in autism. A reported association between autism and deletions of WNT2, a gene also deleted in our patient, is likewise not supported by our case. Previously, fine mapping with microsatellites markers within in a large three-generation family, in which half the members had severe specific language impairment, aided the localization of the SPCH1 locus to 7q31 within markers D7S2459 (107.1 Mb) and D7S643 (120.5 Mb). Additionally, chromosome rearrangement of 7q31 and mutational analyses have supported the growing evidence that FOXP2, a gene within the SPCH1 region, is involved with speech and language development. It is unclear however whether the AUTS1 (autistic spectrum 1) locus, highly linked to 7q31, overlaps with the SPCH1 and FOXP2.

Kevin L. Gunderson

Papers

Power to detect risk alleles using genome-wide tag SNP panels.

Methods for generating amplified nucleic acid arrays

Whole-genome genotyping on bead arrays.

Allele-specific gene expression patterns in primary leukemic cells reveal regulation of gene expression by CpG site methylation

Deletion of 7q31.1 supports involvement of FOXP2 in language impairment: clinical report and review.