The DNA sequence and comparative analysis of human chromosome 20

The finished sequence of human chromosome 20 comprises 59,187,298 base pairs (bp) and represents 99.4% of the euchromatic DNA. A single contig of 26 megabases (Mb) spans the entire short arm, and five contigs separated by gaps totalling 320 kb span the long arm of this metacentric chromosome. An additional 234,339 bp of sequence has been determined within the pericentromeric region of the long arm. We annotated 727 genes and 168 pseudogenes in the sequence. About 64% of these genes have a 5' and a 3' untranslated region and a complete open reading frame. Comparative analysis of the sequence of chromosome 20 to whole-genome shotgun-sequence data of two other vertebrates, the mouse Mus musculus and the puffer fish Tetraodon nigroviridis, provides an independent measure of the efficiency of gene annotation, and indicates that this analysis may account for more than 95% of all coding exons and almost all genes.

The physical maps for sequencing human chromosomes 1, 6, 9, 10, 13, 20 and X

We constructed maps for eight chromosomes (1, 6, 9, 10, 13, 20, X and (previously) 22), representing one-third of the genome, by building landmark maps, isolating bacterial clones and assembling contigs. By this approach, we could establish the long-range organization of the maps early in the project, and all contig extension, gap closure and problem-solving was simplified by containment within local regions. The maps currently represent more than 94% of the euchromatic (gene-containing) regions of these chromosomes in 176 contigs, and contain 96% of the chromosome-specific markers in the human gene map. By measuring the remaining gaps, we can assess chromosome length and coverage in sequenced clones.

A physical map of the human genome

The human genome is by far the largest genome to be sequenced, and its size and complexity present many challenges for sequence assembly. The International Human Genome Sequencing Consortium constructed a map of the whole genome to enable the selection of clones for sequencing and for the accurate assembly of the genome sequence. Here we report the construction of the whole-genome bacterial artificial chromosome (BAC) map and its integration with previous landmark maps and information from mapping efforts focused on specific chromosomal regions. We also describe the integration of sequence data with the map.

Contigs built with fingerprints, markers, and FPC V4.7

Contigs have been assembled, and over 2800 clones selected for sequencing for human chromosomes 9, 10 and 13. Using the FPC (FingerPrinted Contig) software, the contigs are assembled with markers and complete digest fingerprints, and the contigs are ordered and localised by a global framework. Publicly available resources have been used, such as, the 1998 International Gene Map for the framework and the GSC Human BAC fingerprint database for the majority of the fingerprints. Additional markers and fingerprints are generated in-house to supplement this data. To support the scale up of building maps, FPC V4.7 has been extended to use markers with the fingerprints for assembly of contigs, new clones and markers can be automatically added to existing contigs, and poorly assembled contigs are marked accordingly. To test the automatic assembly, a simulated complete digest of 110 Mb of concatenated human sequence was used to create datasets with varying coverage, length of clones, and types of error. When no error was introduced and a tolerance of 7 was used in assembly, the largest contig with no false positive overlaps has 9534 clones with 37 out-of-order clones, that is, the starting coordinates of adjacent clones are in the wrong order. This paper describes the new features in FPC, the scenario for building the maps of chromosomes 9, 10 and 13, and the results from the simulation.

Contigs built with fingerprints, markers, and FPC V4.7

Contigs have been assembled, and over 2800 clones selected for sequencing for human chromosomes 9, 10 and 13. Using the FPC (FingerPrinted Contig) software, the contigs are assembled with markers and complete digest fingerprints, and the contigs are ordered and localised by a global framework. Publicly available resources have been used, such as, the 1998 International Gene Map for the framework and the GSC Human BAC fingerprint database for the majority of the fingerprints. Additional markers and fingerprints are generated in-house to supplement this data. To support the scale up of building maps, FPC V4.7 has been extended to use markers with the fingerprints for assembly of contigs, new clones and markers can be automatically added to existing contigs, and poorly assembled contigs are marked accordingly. To test the automatic assembly, a simulated complete digest of 110 Mb of concatenated human sequence was used to create datasets with varying coverage, length of clones, and types of error. When no error was introduced and a tolerance of 7 was used in assembly, the largest contig with no false positive overlaps has 9534 clones with 37 out-of-order clones, that is, the starting coordinates of adjacent clones are in the wrong order. This paper describes the new features in FPC, the scenario for building the maps of chromosomes 9, 10 and 13, and the results from the simulation.

High-resolution landmark framework for the sequence-ready mapping of Xq23-q26.1

We have established a landmark framework map over 20-25 Mb of the long arm of the human X chromosome using yeast artificial chromosome (YAC) clones. The map has approximately one landmark per 45 kb of DNA and stretches from DXS7531 in proximal Xq23 to DXS895 in proximal Xq26, connecting to published framework maps on its proximal and distal sides. There are three gaps in the framework map resulting from the failure to obtain clone coverage from the YAC resources available. Estimates of the maximum sizes of these gaps have been obtained. The four YAC contigs have been positioned and oriented using somatic-cell hybrids and fluorescence in situ hybridization, and the largest is estimated to cover approximately 15 Mb of DNA. The framework map is being used to assemble a sequence-ready map in large-insert bacterial clones, as part of an international effort to complete the sequence of the X chromosome. PAC and BAC contigs currently cover 18 Mb of the region, and from these, 12 Mb of finished sequence is available.

A physical map of 30,000 human genes

A map of 30,181 human gene-based markers was assembled and integrated with the current genetic map by radiation hybrid mapping. The new gene map contains nearly twice as many genes as the previous release, includes most genes that encode proteins of known function, and is twofold to threefold more accurate than the previous version. A redesigned, more informative and functional World Wide Web site (www.ncbi.nlm.nih.gov/genemap) provides the mapping information and associated data and annotations. This resource constitutes an important infrastructure and tool for the study of complex genetic traits, the positional cloning of disease genes, the cross-referencing of mammalian genomes, and validated human transcribed sequences for large-scale studies of gene expression.

Z extensions to the RHMAPPER package

MOTIVATION

Extensions have been made to the RHMAPPER-1.1 package. One set of extensions computes the totally linked markers and uses the results as input to the salient RHMAPPER functions. The second set of extensions uses TKperl to provide an interactive interface for ease of querying the database and displaying maps.

AVAILABILITY

The extensions can be obtained via ftp.sanger.ac.uk/pub/zmapper.

SUPPLEMENTARY INFORMATION

The User's Manual can be viewed from http:/www.sanger.ac.uk/Users/cari/Z.shtml.

CONTACT

cari@sanger.ac.uk

FPC: a system for building contigs from restriction fingerprinted clones

MOTIVATION

To meet the demands of large-scale sequencing, thousands of clones must be fingerprinted and assembled into contigs. To determine the order of clones, a typical experiment is to digest the clones with one or more restriction enzymes and measure the resulting fragments. The probability of two clones overlapping is based on the similarity of their fragments. A contig contains two or more overlapping clones and a minimal tiling path of clones is selected to be sequenced. Interactive software with algorithmic support is necessary to assemble the clones into contigs quickly.

RESULTS

FPC (fingerprinted contigs) is an interactive program for building contigs from restriction fingerprinted clones. FPC uses an algorithm to cluster clones into contigs based on their probability of coincidence score. For each contig, it builds a consensus band (CB) map which is similar to a restriction map; but it does not try to resolve all the errors. The CB map is used to assign coordinates to the clones based on their alignment to the map and to provide a detailed visualization of the clone overlap. FPC has editing facilities for the user to refine the coordinates and to remove poorly fingerprinted clones. Functions are available for updating an FPC database with new clones. Contigs can easily be merged, split or deleted. Markers can be added to clones and are displayed with the appropriate contig. Sequence-ready clones can be selected and their sequencing status displayed. As such, FPC is an integrated program for the assembly of sequence-ready clones for large-scale sequencing projects.

A gene map of the human genome

The human genome is thought to harbor 50,000 to 100,000 genes, of which about half have been sampled to date in the form of expressed sequence tags. An international consortium was organized to develop and map gene-based sequence tagged site markers on a set of two radiation hybrid panels and a yeast artificial chromosome library. More than 16,000 human genes have been mapped relative to a framework map that contains about 1000 polymorphic genetic markers. The gene map unifies the existing genetic and physical maps with the nucleotide and protein sequence databases in a fashion that should speed the discovery of genes underlying inherited human disease. The integrated resource is available through a site on the World Wide Web at http://www.ncbi.nlm.nih.gov/SCIENCE96/.

SAM: a system for iteratively building marker maps

SAM (system for assembling markers) is a system which supports man-machine problem solving for iteratively ordering a set of markers. SAM aids the user in partially ordering a set of markers based on incomplete and uncertain data. As data is added and modified, SAM aids the user in updating the previously assembled maps. The input is a file of clones and for each clone, a list of the markers contained within it. The objective is to order the set of markers such that the markers contained in each clone are consecutive. The user directs the map building by selecting functions to assemble a region of markers, order the clones to fit the order of the markers and position new markers within an ordered set of markers. The user can edit the input data, edit the assembled map and add clones to the map based on their marker content. The results are displayed graphically and can be saved in a solution file. Based on the partial map, the user designs new experiments or edits the existing data to fill gaps and resolve ambiguities. When a previously assembled map is loaded into SAM, it is automatically updated with the new or altered data. SAM treats all markers as points, but has special features for multiple copy and long markers so that they can be used in the map building process. This system has supported the building of a YAC map of human chromosome 22 at the Sanger Centre, where use of Alu-PCR product markers is a major component in determining clone overlap and where we have an on-going effort to accumulate data from various sources. SAM is also being used at various other laboratories.

GRAM and genfragII: solving and testing the single-digest, partially ordered restriction map problem

GRAM (Genomic Restriction map AsseMbly) takes as input single-digest restriction fragments for a set of overlapping clones and outputs one or more plausible partially ordered restriction maps. For each restriction map, GRAM shows the corresponding alignment of the input clone fragments. Due to the error and uncertainty in experimental data, this problem is computationally difficult to solve; therefore, the principle objective in the design of GRAM is to facilitate man-machine collaborative problem solving. GRAM quickly approximates a solution, as follows. (i) A clustering algorithm determines a probable set of restriction fragments. (ii) An assembly algorithm permutes the set of restriction fragments such that the maximal number of clone fragments are contiguous. The output of the GRAM algorithm is displayed for the user to query and edit. This paper describes the stochastic assembly algorithm and shows how it works with the interactive graphics to support man-machine problem solving. In order to test and verify the performance of GRAM, we have developed a program called genfragII to simulate the digestion of clones and fragments; this program is described and results are presented. GRAM is also being used for a number of genome mapping projects.

Information contents and dinucleotide compositions of plant intron sequences vary with evolutionary origin

The DNA sequence composition of 526 dicot and 345 monocot intron sequences have been characterized using computational methods. Splice site information content and bulk intron and exon dinucleotide composition were determined. Positions 4 and 5 of 5' splice sites contain different statistically significant levels of information in the two groups. Basal levels of information in introns are higher in dicots than in monocots. Two dinucleotide groups, WW (AA, AU, UA, UU) and SS (CC, CG, GC, GG) have significantly different frequencies in exons and introns of the two plant groups. These results suggest that the mechanisms of splice-site recognition and binding may differ between dicot and monocot plants.