Abstract
OBJECTIVE To analyze coding SNPs of the HLA-DQA1 gene involved in susceptibility for cervical cancer by a bioinformatics approach, and to choose some SNPs that may have an association with cervical cancer.
METHODS By a SNPper tool we extracted SNPs from a public database (dbSNP), exporting them in FASTA formats suitable for subsequent use. Then we used PARSESNP as a tool for the analysis of the cSNPs.
RESULTS In the cSNPs of the HLA-DQA1 gene, we find that rs9272693 and rs9272703, are made up of missense mutations which convert a codon for one amino acid into a codon for a different amino acid. We chose a PSSM Difference >10 as a lower level for the scores of changes predicted to be deldterious.
CONCLUSION We used a bioinformatics approach for cSNPs analysis of the HLA-DQA1 gene. This method can select the variants in a conserved region, and give a PSSM Difference score. But the results need to be verified in cervical cancer patients and a control population.
keywords
Cervical cancer is the third most common cancer in women worldwide.[1] Infection with oncogenic types of human papillomavirus (HPV) is the main cause of cervical cancer and its precursor lesions [cervical intraepithelial neoplasia (CIN)]. During their life-time many women become infected with HPV, but only a minority develop CIN or cervical cancer. Consequently, there have to be other factors, e.g., genetic factors, that play a role in the development of CIN or cervical cancer. Almost all research on cervical cancer susceptibility has focused on genes in the HLA-complex. The HLA-complex, on the short arm of chromosome 6 (6p21.3), contains Class-I, -II, -III and other 200 genes with known or unknown functions and a strong LD exists between them. The function of both Class-I and Class -II genes is the presentation of short, pathogen-derived peptides to T cells. The products of the Class-I genes (HLA-A, -B and -C) are usually associated with presentation of endogenous proteins. Class-II genes (HLA-DR, -DQ, and -DP) are associated with presentation of exogenous proteins.[2] Zoodsma et al [3] identified all published studies from 1980 to January 2002 on the PubMed databases. They focused on common and genetic risk factors such as HLA and other genes (Tp53, IL-10, CYP2D6 and MTHFR) that may be involved in susceptibility to (pre) neoplastic cervical disease. We selected HLA-DQA1 for further analysis.
Single nucleotide polymorphisms (SNPs) are an increasingly important resource for understanding the structure and history of the human genome. A SNP is defined as a mutation involving a single DNA base substitution that is observed with a frequency of at least 1% in a given population. SNPs are the most common form of genetic variation in humans. Overall, SNPs account for 90% of inter-individual variability. [4] Scientific advancements have resulted in a series of genetic markers with ever-increasing information content and resolving power. In the past 30 years, restriction fragment length polymorphisms (RFLPs), short tandem repeats (STRs), and SNPs have played significant roles in genetic research. In order to analyze coding SNPs (cSNPs) in the HLA-DQA1 gene, which is a subtype of HLA-II, we planned to devise a platform to choose cSNPs by bioinformatics tools and predict variation that is likely to have a functional effect.
MATERIALS AND METHODS
Retrieval of human cSNPs for the HLA-DQA1 gene
SNPper is a web-based tool to automate the task of extracting SNPs from public databases, to analyze them and to export them in formats suitable for subsequent use.[5] SNPper is freely available at http://sņpper.chip.org/. Registration is optional and it provides access to some advanced features. The most important public SNP database is dbSNP (accessible at http://www.ncbi.nih.gov/SNP/), that collects all SNPs detected by either computational methods (i.e. comparing matching sequences stored in databases like GenBank) or direct observation. The general purpose of SNPper is to create sets of SNPs responding to user-defined criteria. SNPs can be retrieved through their name. On the SNP per’s interface we selected “Gene Finder”, input gene’s Symbol and then retrieved all the SNPs for the HLA-DQA1 gene. SNPper allows the user to filter or refine a SNP set to show the the cSNPs of the HLA-DQA1 gene. Lastly we can export the information that SNPper associates to each SNP.
Extracting FASTA sequences of human cSNPs for the HLA-DQAl gene
FASTA sequences of human cSNPs for the HLA-DQA1 gene were retrieved from public database db-SNP(accessible at http://www.ncbi.mfrgov/SNP/). One can submit all the SNPrs#, then the web will automatically send the FASTA sequences to your E-mail within 24 h.
Searching for homology models of the HLA-DQAl gene
In order to predict the severity of the effect of a missense change on function, we had to obtain the homology models. PARSESNP accepts blocks. Blocks are multiply aligned ungapped segments corresponding to the most highly conserved regions of proteins. Information about Blocks is available at http://blocks.fhcrc.org/o We used Reverse-PSI BLAST Searcher as a tool to search for homology models of the HLA-DQAl gene/[6]
Analyzing for cSNPs of the HLA-DQAl gene
PARSESNP is a tool to display and analyze polymorphisms in genes and is available on the World Wide Web at http://www.prowgb.org/parsesnp/[7] Using a reference DNA sequence, an exon/intron position model and a list of polymorphisms, that information can be extracted from GenBank (http://ncbi.nlm.mh.gov). It determines the effects of these polymorphisms on the expressed gene product, as well as the changes in restriction enzyme recognition sites. The results were saved in our computer.
RESULTS
Results of human cSNPs for the HLA-DQAl gene
Table 1 shows the SNP set export form of the HLA-DQAl gene. It displays the SNP rs#, SNP position, band, distance from previous SNP, alleles, gene and role. From the column of distance from previous SNP, we can see that the density of SNPs is high at the chromosome 6p21.32 band, far from the average distance that SNPs occur at approximately every 1000 bases over the human population.
The SNP set export form of the HLA-DQA1 gene
Result of homology models for the HLA-DQAl gene
We looked for Blocks in the Blocks Database at the http://www.block.fhcrc.org The result is IPB001003,Class II histocompatibility antigen, alpha chain, alpha-1 domain, Score 382 bits and E Value e107.
Analysis results for cSNPs of the HLA-DQAl gene
The results of PARSESNP can be viewed in a variety of different formats (Figs. 1, 2, Table 2). Out of the 50 SNPs in the resulting database by SNPper, only 14 SNPs showed analytic results by PARSESNP, the other 36 SNPs showed no BLAST results on the genomic sequence or on the coding sequence. Fig.l displays the locations of the polymorphisms in the gene (both coding sequence and genomic sequence). Table 2 describes in detail the effect of each polymorphism in the gene, including nucleotide change, amino acid result, restriction enzyme polymorphisms and PSSM difference. If the region containing a missense change is aligned to a block, one can attempt to gauge the effect of the change by examining the change in the PSSM score. We have chosen 10 as a rough lower level for the scores of changes predicted to be deleterious. Some nucleotide substitutions do not represent a change in the encoded amino acid, and are termed ‘synonymous’ cSNPs. Non·synonymous cSNPs can be those that result in conservative substitutions (amino acids with a similar size or charge) or non-conservative substitutions. Of the 14 SNPs, were 85.7/o for non-synonymous cSNPs, and 14.3% for synonymous cSNPs. In Table 2 rs9272693 and rs9272703 belong to non-synonymous cSNPs with a PSSM Difference >10 and SIFT Score <0.05. These have been empirically determined to be deleterious. Fig.2 displays the effect of the change form from the original amino acid, amino acid position, and new amino acid (*for stop codon). The variant of rs9272693 is the first nucleotide change C to T, which lead to the amino acid change from Arg to Trp. The variant of rs9272703 is the second nucleotide change G to T, which lead to the amino acid change of Gly to Cys.
Polymorphisms in the sequence. Fig.2 Polymorphisms in the coding sequence from Fig.1 and Table 2. This region shows a Block aligned in the coding sequence. Two missense changes (number 1 and 5 in Table 2) in red and the other missense changes in black. Two missense changes (number 1 and 5) are colored red because the PSSM difference s∞re ¡s>10. The residues of the reference protein are colored to indicate how each position compares to the aligned Block. Those residues ∞lored green are most similar to the correspong¡ng column in the Block, while those colored red are most diverged.
Table of polymorphisms for the HLA-DQA1 gene
Overview of polymorphisms and Blocks in the HLA-DQA1 gene, (a) Genomic Sequence (b) Coding Sequence. Overview of polymorphisms and blocks in the HLA-DQA1 gene. The sequence and variants ∞me from GenBank NC_000006.9 and the homology model deπves from the Block IPB001003 families. In the ‘Genomic Sequence’ plot, the top region of the graphics shows the locations of the Blocks on the reference sequence. The green graphics correspond to blocks IPB001003A, IPB001003B and IPB001003C. The middle shows the locations of the exons, represented by boxes. The bottom region displays the locations of the polymorphisms, the first row displays missense changes in black and the se∞nd row shows silent changes in purple.
DISCUSSION
The technical term, SNPs, appeared in human molecular genetic literature for the first time in 1994. SNPs are tightly associated with complex diseases. Association studies try to establish a relationship between a phenotype and one or more regions of the genome and the distribution and function of SNPs are important areas of current research. A variant may affect the expression or translation of a gene product, either by interrupting a regulatory region or by interfering with normal splicing and mRNA function. This can include SNPs in regulatory SNPs, intronic SNPs and exon-intron boundary SNPs. Research suggests that most SNPs fall in the 95 percent non-coding region of the genome with only 5 percent falling in the coding region. [8] Non-synonymous SNPs alter the amino acid substitution or introduction of a nonsense/truncation mutation.[9] The main purpose of this study was to predict the severity of the effect of a missense change on function. To assess the possible damaging effect of amino acid substitution, we developed a bioinformatics platform to analyze coding SNPs of the HLA-DQAI gene involved in susceptibility for cervical cancer.
Today, the primary database of polymorphisms is dbSNP, which currently contains more than 5,000,000 validated human SNPs. A powerful resource for SNP analysis is SNPper. SNPper was created in the Kohane Lab at Harvard University for the analysis of SNPs. SNPper focuses on SNP selection for genetic studies and is freely available. Mooney’[10] showed that general disease-associated mutations tend to occur in positions that are conserved. PARSESNP is a tool for the display and analysis of polymorphisms in genes. In order to assess the effect missense changes have on gene product function, PARSESNP prorides a method of submitting homology information. PARSESNP accepts Blocks, a format that represents distinct regions of ungapped alignment in protein sequences from the Blocks database. The severity of the effect of a missense change on function can be predicted using the homology models.
There were 1446 SNPs in the HLA-DQAl gene extracted by SNPper. The average distance was 17 nucleotides, which indicates that the density in this gene is very high. On coding sequence there are 50 SNPs, 14 SNPs of those produced analytic results by PARSESNP, another 36 SNPs could not get any BLAST results in the Genomic Sequence or coding sequence. A large number of redundant, incomplete, even incorrect SNPs are also collected in the SNP databases because of various resources of SNPs, such as: the results of sequencing, the BLAST of EST, variation in the results of experiments and so on. [11] Among the 14 SNPs which had analytic results, the PSSM Differences of rs9272693 and rs9272703 were more than 10. As for the prognosis of detrimental mutations, the probability of a deleterious variant is large if PSSM Difference >10.[7] The biologic significance is that the amino acid sequence coded by the variant nucleotide of rs9272693 and rs9272703 is changed, which probably alters the function of the HLA-DQAl gene in production of exogenous proteins, and thus changes the immune reaction of an individual to HPV, and finally the susceptibility of cervical cancer increases.
In 2001, 1,420,000 SNPs in the human genome were reported in Nature by the International SNP Research Association and International Human Genome Sequencing Association. The data showed that SNPs occur approximately every 1250 bases. Up until the present, the number of known SNPs has grown at an ever increasing rate. After the Human Genome Project the study of SNPs has been a new focus of international research. Preliminary studies indicate that there are obvious differences between the Chinese and Western populations in the frequency of SNPs in several important diseases. Abundant hereditary resources in our country should be utilized to conduct SNPs research on important diseases with particular emphasis on constructing a genome SNPs’ systematic catalogue of Chinese people. This would be quite meaningful for future health care and the biotechnologic medical industry. The number of SNPs is numerous and more than half are nonsense mutations,[10] so we need to increase the degree of success of experimental tests using bioinformatics tools to screen SNPs of people whose phenotypes and functions are altered. The solving of the problem depends on the development of the software and researchers’ complete understanding and mastering of the software.
In this article we offer a set of practical, feasible approaches to solve the problem. The HLA-DQAl gene studied in this report is based on previous research on cervical cancer. The mutations of the HLA-DQAl gene may alter the function of the gene, and reduce the immune response of patients to HPV infection resulting in the promotion of cervical cancer. However this hypothesis needs further studies by measuring the frequencies of a number of SNPs in two populations, and by detecting SNPs that show a significant difference in frequency.
- Received September 30, 2005.
- Accepted November 25, 2005.
- Copyright © 2006 by Tianjin Medical University Cancer Institute & Hospital and Springer
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.