Top

  Background

To date, GeneCards, like many other major gene indices including its two major resourcesLocusLink and Ensembl, is based mainly on full length mRNA sequences and genes predicted from genomic data. Since sequencing full length mRNA is both time consuming and costly, the mRNA sequences for many genes are not yet available, which results in the absence of numerous genes from the gene index.

Since the early 1990's high throughput methods have generated a substantial number (>5 million) of Expressed Sequence Tags (ESTs), which now offer the most extensive window to the entire human transcriptome, and to the genes coded within it. Unfortunately, given their fragmentary nature (typically 400-600 bases) and inaccurate information (1-3% sequencing errors), assigning each of these ESTs to genes has been elusive. Previous and ongoing projects designed to address this problem have resulted in various gene lists that exhibit only partial overlap.

Top

  Goals

GeneTIDE aims to integrate various data resources in order to create a comprehensive list of human genes. This is done by association between the set of over ~5.5 million human ESTs currently available from dbEST and mRNA sequences from GenBank to the set of ~35,000 human genes as defined in GeneCards. Heretofore transcripts (mRNA & EST) can be :

1. Proven to belong to an existing GeneCards gene

2. Used to define de-novo genes

3. Demonstrated to be an artifact or contaminated(genomic DNA, vector, etc.), and should therefore be discarded.

Top

  Methods

  GeneTIDE's generation process consists of two major stages.

1. Association of transcripts with existing GeneCards genes. This is done by integration of data from several resources.
   • Resources :
     • UniGene - UniGene is an experimental system for automatically partitioning GenBank sequences into a non-redundant set of gene-oriented clusters. Some of these UniGene clusters are associated with LocusLink gene identifiers, which are in turn connected to a specific GeneCards gene. This gene is therefore applied to all the transcript members of the cluster.

     • DoTS - DoTS (Database Of Transcribed Sequences) is a human and mouse transcript index created from all publicly available transcript sequences. The input sequences are clustered and assembled to form the DoTS Consensus Transcripts (dots) that comprise the index. Similarly to UniGene clusters, the "dots" are associated with a Locuslnk gene and hence a GeneCards gene.

     • AceView - AceView offers an integrated view of the human and nematode genes as reconstructed by alignment of all publicly available mRNAs and ESTs on the genome sequence. The asscoation of each transcript with a gene is extracted and incroporated into GeneTIDE.

     • GeneAnnot - GeneAnnot is an in-house automated system for annotation of Affymetrix HG-U95 probe sets, which is revised and improved by direct sequence comparison of probes to GenBank, RefSeq and Ensembl mRNA sequences. As probe sets are dervied from transcripts (mRNA or EST) the annotation given by GeneAnnot to a specific probe set, is applied in GeneTIDE to the transcript from which it was derived.

     • GeneLoc - Genomic locations for the transcripts in question, which were obtained using BLAT were downloaded from UCSC's genome browser database and compared with data from GeneLoc, an exon-based system which forms part of the GeneCards suite of databases and integrates data from LocusLink and Ensembl to create a unified location for each gene. The genes that are located on the same genomic region as found by BLAT for a specific EST were recorded.

   • Scoring scheme :

     The various gene annotations obtained for each transcript by the five aforementioned methods (UniGene, DoTS, AceView, BLAT & GeneLoc, and GeneAnnot) are integrated into a single scoring scheme which is composed of two major parameters each ranging between 0 and 1. Consensus (Co) and Uniqueness (Uq) are defined for a pair (E,G), where E is an EST and G is a GeneCards gene. A third value, score, summarizes these two parameters into one. For a more detailed view on how these parameters are calculated click here.

2. Definition of new transcript clusters as a basis for new genes Previously unannotated remaining transcripts are subjected to a second stage of analysis. In order to define groups of transcripts that form a new gene, clustering data from UniGene, DoTS and AceView is used. Groups of transcripts (ESTs & mRNA), with more than 1 transcript (for quality consideraions), for which every member transcript belongs to the same UniGene, DoTS and AceView clusters are defined as an EST based Gene Candidate (EGCs).

   EGCs are annotated with various parameters that help determine a probabilty level for whether they truly form a new, previously undefined, gene. Alternatively, an EGC may be shown to be a representative of an exon previously unknown (e.g. due to alternative splicing) of an existing gene. The annotation includes :

   • Identifiers of the UniGene, DoTS and Aceview clusters which contain the EGC's transcripts
   • Genomic location of each transcript
   • Indication of whether the transcripts are spliced or not
   • Indication of whether the transcripts include sequences that were detected by RepeatMasker as contaminated
   • Expression vectors according to GeneNote data of probe sets derived from this EGC's member transcripts
   • The identity and distance (in base pairs) of the GeneCards gene closest to the majority of member transcripts

Top

  Searching GeneTIDE

GeneTIDE is a transcript centered database. Therefore, most annotation data is given from a transcript point of view.

  Simple query

Top

  Multiple Sequences Queries

In order to use batch queries either upload a file containing GenBank IDs (see example) or simply paste the GenBank IDs into the query window. After submitting the multiple sequences query you will get a link to the result file which can be downloaded for your use. Click here for more details on the output file format.

Top

  EGC List

1. EGC - The name of the EGC.
2. Size - Number of transcripts contained within the EGC.
3. RepMask - The number of transcripts within the EGC found by RepeatMasker to contain repetitive elments.
4. Expressed - Number of transcripts that had probe sets derived from them, and shown according to GeneNote data to have an expression pattern.
5. Spliced - The number of transcripts, among the EGC's members, which have a known splice site.
6. CNG_GC_ID - Common Nearest Gene (CNG) is the gene that has the greatest number of transcripts among the EGC's members that this gene is their nearest gene. Here the GC ID of the gene is displayed.
7. CNG_Symbol - The gene name (or symbol) of the Common Nearest Gene.
8. CNG_Symbol - The gene name (or symbol) of the Common Nearest Gene.
9. CNG_distance - The average distance (in bp), of transcripts for which the Common Nearest Gene is their nearest gene, from that gene.

Top

  Interpreting the results

Query results will be displayed in accordance with the search key used :

• U95 Probe Set
• U133 Probe Set
• GeneCards gene
• GeneTIDE EGC number
• UniGene Hs. number
• DoTS transcript

Top

  General links

The first table offers links to commonly requested EST related data :

• Similar transcripts - other transcripts that share the same Unigene, DoTS and AceView clusters as the queried transcript.
• FASTA sequence - The FASTA format sequence of the transcript, retrieved from NCBI.
• UCSC genome broswer - The genomic area of the transcript diplayed on the UCSC genome broswer along with the "GeneCards Genes" custom track.

Top

  Validation Level

An indication to the probability that this is avalid transcript and not an artifact or contamination. This value ranges from 0 to 4, corresponding to the following verbal descriptions :

• 0 - Very low
• 1 - Low
• 2 - Medium
• 3 - High
• 4 - Very High

The validation level is caclulated in the following manner. Each of the three resource (UniGene, DoTS and AceView) that has not filtered this transcript out of its analysis earns the transcript one point. Next, RepeatMasker was applied to the transcript's genomic location. If this area is announced contaminated by any repetitive element by RepeatMasker, the transcript loses an additional point. The overall highest (most valid) score is 4.

Top

  Resources Annotation

A summary table showing the gene association given by each of the five resources (UniGene, DoTs, AceView, Geneannot, BLAT&GeneLoc) to this transcript.

• UniGene - The UniGene cluster the transcript is contained in, the LocusLink ID associated with this cluster according to UniGene, and the GeneCards GC ID and symbol of this gene

• DoTS - The DoTS cluster the transcript is contained in, the LocusLink ID associated with this cluster according to DoTS, and the GeneCards GC ID and symbol of this gene

• AceView - The AceView gene the transcript is contained in, the LocusLink ID associated with this gene according to AceView, and the GeneCards GC ID and symbol of this gene

• GeneAnnot - The Affymetrix U95 probe set derived from this transcript, the quality of annotation, and the GeneCards GC ID and symbol of the genes associated with this probe set.

• BLAT & GeneLoc - Associates the transcript with genes that have exons at the same genomic location as the transcript. The data includes the GeneCards GC ID and symbol of the associated gene, and confirmation according to LocusLink and Ensembl of this annotation (+ confirmed, - not confirmed).

Top

  Genomic location based data

Information regarding the queried gene acquired by aligning it against the genome using BLAT.

• Genomic Location - Chromosome, strand, start & stop coordinates of the genomic locations where the transcript is aligned to according to BLAT. The orientation field expresses the result of running UCSC's polyinfo program on this transcript, which trys to determine the most probable orientation of the transcript. Possible values are 'alignment' for a result consistent with the value suggested in the strand field, 'opposite' for the opposite strand than the one given in the strand field, and 'unknown' for an inconclusive result.

• Gene Territory / Nearest Gene - GeneCards genes which the transcript resides within their genomic territory according to GeneLoc (not to be confused with the 'BLAT & GeneLoc' resource annotation section, since the transcript does not necessarily intersect exons, and could reside in an intron to be included here ). If the transcript does not reside within the genomic territory of any gene, the nearest gene and its distance in bp is displayed.

• Splicing - Indicates whether this transcript undergoes splicing or not. This data is retrieved from UCSC's genome browser 'Spliced ESTs' track.

• RepeatMasker - Indicates if there are any areas of the transcript found to be repetitive elements according to the RepeatMasker program. This includes the chromosome, strand and cooridnates of the genomic location (onto which the transcript was aligned) where the repetitive element resides as well as the type of the reptitive elemene.

Top

  Gene Association -

The results summary displays the "bottom line" of assocations between the transcript in question and existing GeneCards genes. The table shows all genes that were associated with the transcript via any of the resources. This is followed by an indication of which resources support the specific association, and the Concensus(Co) and Uniqueness(Uq) values that summarize them. Score is a value ranging between 0 and 1 that recapitualtes the previous two values into one. Rank is the position of the specific gene among all other genes assoicated with this transcript. The ranks are sorted according to score and Uniqueness(Uq).

  Contamination level -

The probability that this transcript is contaminated is given in terms of 3 categoriesn according to the contmaination score calculated earlier :

• Not contaminated : Contamination levels 0-1
• Inconclusive : Contamination level 2
• Contaminated : Contamination levels 3-4

  U133 Probe Set

  GeneCards gene

  TIDE EGC number

Top

  EGC gene properties

The first table offers links to the UniGene, DoTS and AceView clusters from which this EGC was constructed.

Top

  Member transcripts

A list of transcripts constituting this EGC. For each tracnscript we show :

• Transcript type - EST or full length mRNA.
• Genomic location - Chromosome, strand and coordinates.
• Splicing - Is the trnascript spliced or not.
• RepeatMasker - Is the transcript contaminated by a repetitive element, according to RepeatMasker.

Top

  Expression patterns

If any of the transcripts withing this EGC had Affymetrix U95 GeneChip probe sets derived from them, the expression pattern of these probe sets across 12 normal human tissues according to GeneNote data will be displayed here. For more information regarding GeneNote click here.

Top

  Common Nearest Gene

This field shows the identity of the GeneCards gene which for biggest number of transcripts from this EGC, is their nearest gene.

  UniGene Hs. number

  DoTS transcript

Displays a short description of the requested 'DoTS Transcript' cluster, and the GenBank accessions contained within it, along with hyperlinks to the indiviual GeneTIDE pages of these transcripts.

  Multiple Sequences Queries

genbank_id	gc_id	gene_symbol	Co	Uq	score	rank
BE874217	GC01M019135	CAPZB	0.6	0.833333	0.726101	1
BE874217	GC18M043619	MADH2	0.2	0.5	0.380789	2
#########EGC#############
AA001179	EGC6762
AA001185	EGC5458
AA001199	EGC9493
AA001205	EGC9460
#########UNKNOWN#########
DE874217

The first section, Gene associations is equivalent to the result obtained from the integrated scores table that is created for querying each single GenBank ID individually. In the second section, EGC transcripts that belong to a newly defined EST based GeneCards Candidate (EGC) are displayed in the first column, and the EGC identifier is displayed in the second column. In the third section, Unknown, appears a list of queried GeneBank IDS that were neither associated to a known gene nor belong to an EGC.

Top

  About the author

GeneTide was created by Maxim Shklar (maxim.shklar@weizmann.ac.il , shklar@bigfoot.com) during 2003-2004, as part of the thesis for a M.Sc degree in Life Science - Bioinformatics at the Feinberg Graduate School , Weizmann Institute of Science, Rehovot, Israel.

  Current GeneTide team

GeneTide is currently maintained by Shany Ron, Tsippi Iny Stein and Ohad Greenshpan from the
GeneCards team.
GeneTide is a member of the GeneCards suite of databases.

Back to the top

Back to GeneTide Home page