Bioinformatics Advance Access originally published online on July 5, 2005
Bioinformatics 2005 21(17):3469-3474; doi:10.1093/bioinformatics/bti566
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
A precise and scalable method for querying genes in chromosomal banding regions based on cytogenetic annotations
Department of Computer Science and Information Engineering, National Cheng Kung University No 1. Da-Shueh Road Tainan, Taiwan
1Department of Microbiology and Immunology, College of Medicine, National Cheng Kung University No 1. Da-Shueh Road Tainan, Taiwan
2Department of Pathology, National Cheng Kung University No 1. Da-Shueh Road Tainan, Taiwan
3Institute of Molecular Medicine, National Cheng Kung University No 1. Da-Shueh Road Tainan, Taiwan
*To whom correspondence should be addressed.
 |
Abstract |
|---|
 TOP Abstract 1 INTRODUCTION2 SYSTEM AND METHODS3 IMPLEMENTATION4 PERFORMANCE EVALUATION5 DISCUSSIONS AND CONCLUSIONSREFERENCES |
|---|
Motivation: Staining the human metaphase chromosomes reveals characteristic banding patterns known as cytogenetic bands or cytobands. Using technologies based on metaphase chromosomes, researchers have accumulated much knowledge about the correlations between human diseases and specific cytoband aberrations, indicating the presence of disease-associated genes in those bands. With the progress of human genome project and techniques such as fluorescent in situ hybridization, many genes have been assigned to the cytobands and annotated in public databases, making it possible to find all genes in the disease-related cytobands through database queries. However, finding genes in cytobands remains an imprecise process, partly due to the insufficiency of current methods for cytoband queries, especially for those based on cytogenetic annotations.
Results: By transforming the cytoband annotations into numerical segments, a new query method is developed that is able to accurately define any cytogenetic ranges in human chromosomes. A query system (designated cytoband query sys CQS) is implemented using cytogenetic annotations in the public domain. Judged by a performance test, CQS executed as accurately as expected using cytogenetic annotations from NCBI Map Viewer. The new method is scalable and can be applied to genomes from other species.
Availability: The CQS is freely accessible over the Internet at http://moris.csie.ncku.edu.tw/cqs/
Contact: clh9{at}mail.ncku.edu.tw
Supplementary information: http://moris.csie.ncku.edu.tw/cqs/
|
1 INTRODUCTION |
|---|
|
TOPAbstract1 INTRODUCTION2 SYSTEM AND METHODS3 IMPLEMENTATION4 PERFORMANCE EVALUATION5 DISCUSSIONS AND CONCLUSIONSREFERENCES |
|---|
In humans, there are 22 pairs of autosomal chromosomes (chromosomes 1–22) and a pair of sex chromosomes (XX for female and XY for male). In the late 1960s, researchers found that quinacrine mustard stained the chromosomes of plants and produced characteristic staining patterns (Caspersson et al., 1968). This banding technique was soon applied to human chromosomes (Caspersson et al., 1971) and it turned out that the banding patterns are detailed and reproducible enough that not only normal chromosomes can be readily identified without confusion, but major structural aberrations can also be characterized. Since then, many disease-specific major structural aberrations were described in the literature (Mitelman and Heim, 1988). With metaphase chromosomal techniques like G-banding, Comparative Genomic Hybridization (Kallioniemi et al., 1992), and Spectral Karyotyping (Liyanage et al., 1996; Schrock et al., 1996), a lot of non-random chromosomal aberrations were shown to correlate to the diagnoses, prognoses and treatment responses of many human diseases (Shikoshi et al., 1994; Weinstein et al., 1969; Zech et al., 1976). Although tracing these gross abnormalities to the gene level has revealed the molecular mechanism of many diseases (Collins, 1995), far more genes remain to be identified whose alterations are the key events in the non-random chromosomal aberrations. There are several methods to look for the key molecules. One approach starts with the identification of all genes in the cytogenetic bands involved in the chromosomal aberrations.
Many genes have been mapped to cytogenetic bands by fluorescent in situ hybridization (FISH), a technique to visualize the locations of genes on the chromosomes (Kirsch et al., 2000; Korenberg et al., 1999). The human genome project also helps to annotate the cytogenetic positions of genes. However, even though many genes have been assigned to cytogenetic bands, there is no precise method based on such information that is able to find all genes in every user-defined cytogenetic query region. Such a deficiency is due to the peculiarity of the nomenclature system for cytogenetic bands and the failure of current methods to solve the problems. One of the methods is string search. The usefulness of string search is very limited because it cannot handle ranges of cytogenetic bands. For example, band 8q23 is included in the region 8q22–8q24. However, using the string-search method, the query ‘8q23’ will not find ‘8q22–8q24’ because the string ‘8q23’ is not present in ‘8q22–8q24’.
In this study, we propose an approach to find genes in cytobands, in which cytogenetic banding regions are transformed into numerical segments by themselves without mapping to other numerical systems such as sequence positions. In this way, every cytoband data or query string becomes a chromosome number and a numerical segment with a start value and a stop value. By comparing these values, a complicated cytoband query (CQ) will be transformed into a simple numerical calculation and set operation. As a consequence, correct results can be retrieved quickly and precisely by a simple query to the database.
In the following sections, we will describe the model of CQ, the mapping function of cytoband annotation to the number system, and the implementation of the cytoband query system (CQS). The performance of CQS will be evaluated and compared to NCBI Map Viewer, a cytoband related project supported by the National Center for Biotechnology Information (NCBI).
|
2 SYSTEM AND METHODS |
|---|
|
TOPAbstract1 INTRODUCTION2 SYSTEM AND METHODS3 IMPLEMENTATION4 PERFORMANCE EVALUATION5 DISCUSSIONS AND CONCLUSIONSREFERENCES |
|---|
The solution of CQ will be presented in this section. The model of CQ will be described first. Secondly, the nomenclature system of cytobands will be discussed, which is the main reason why string comparison cannot solve the CQ problem. Thirdly, the method to transform the cytoband annotations to a numerical system will be demonstrated. This method ensures that every query will give rise to precise results.
2.1 The model of CQ
We call the scope of one cytogenetic band or the scope from one cytogenetic band to another cytogenetic band a cytogenetic banding region. The CQS wants to find all the genes in the database whose cytogenetic annotations are relevant to a query region. All cytogenetic annotations in the database can be classified into six groups by their relationships to a query region as illustrated in Fig. 1. The first line stands for a query region, while the rest represent all the cytogenetic annotations of genes in the database. We may see that groups 5 and 6 have no intersections with the query region and therefore are irrelevant. The remaining groups 1–4 contain genes that researchers may be interested in. In groups 1 and 4, the query region and the cytogenetic annotations intersect on one side. Group 2 genes reside in the query region while group 3 genes include the query region. In general, group 2 genes are what researchers want most from a CQ, followed by groups 1, 4 and 3. Our aim is to design a precise and robust method that can find all 6 groups of cytoband annotations relating to a query region. In order to understand the method, we will explain the cytoband nomenclature first.
|
2.2 Cytoband nomenclature
Each chromosome contains a centromere which divides the chromosome into a short arm and a long arm, designated as p and q, respectively. Regions and bands are numbered successively from the centromere outward along each chromosome arm. Certain conspicuous banding landmarks are chosen to divide each arm into regions. Such a landmark is considered to belong entirely to the region distal to the landmark and is the first band in that region. Therefore, starting from the centromere and going toward the telomere, the first chosen landmark encountered would be the beginning of region 2; the next, more distal landmark would become the first band of region 3; and so it goes on for all chosen landmarks. The two regions adjacent to the centromere are assigned to be region 1 in each arm. The regions thus defined can be further divided into two or more bands according to the staining patterns in the regions. For example, the region 8q2 is divided into four bands (Fig. 2): 8q21 (the landmark), 8q22, 8q23 and 8q24. With increased banding resolution by using chromosomes at the late prophase, some of the low resolution bands in metaphase chromosomes could be further divided by the newly appearing high resolution bands. When an existing band is subdivided, a decimal point is placed behind the original band designation and is followed by the number assigned to the sub-bands. For example, a gray-staining, high-resolution band subdivides band 8q22 into sub-bands 8q22.1, 8q22.2 and 8q22.3, with 8q22.1 closer to the centromere and 8q22.3 closer to the telomere. If a sub-band were further subdivided, additional digits will be added to the designation. A band is usually subdivided into three sub-bands, even though theoretically it can be subdivided into any number of new bands. The centromere is labeled as ‘cen’, sometimes designated as p10 or q10. Telomeres reside at the ends of the chromosomal arms. They are labeled as ‘ter’. In summary, the complete designation of a chromosomal band is composed of: (1) the chromosome number, (2) the arm symbol, (3) the region number, (4) the band number within that region and optionally (5) a decimal point followed by the sub-band number.
|
2.3 Mapping a cytoband region to an integer interval
Our method transforms the query region and all cytoband annotations in the database into integer intervals. The problem of CQ is thus simplified so as to find every integer interval in the database that is relevant to the query region, i.e. groups 1–4 in Fig. 1. We designate this system as CQS. The mapping function is described in the following.
From the viewpoint of cytoband nomenclature, every cytoband is expressed as s[p, q]s1s2.s3s4s5, where s is the chromosome number, the symbol [p, q] means the p-arm or q-arm of that chromosome, s1 the region number, s2 the band number within s1, and s3–s5 the sub-band numbers. In the expression of a cytoband the symbols p, q, s1, s2, s3, s4 and s5 are in order from large scope to small scope, and these symbols may be absent in inverse order.
In the expression of cytoband s[p, q]s1s2.s3s4s5, we define that si = 0, if si is absent from the expression. Otherwise, si is a nature number between 1 and 9. Let p-arm be mapped to negative integers, q-arm to positive integers and ‘cen’ to zero. As telomeres reside at the ends of a chromosomal arm, we let ‘pter’ be mapped to the left-end and ‘qter’ maps to the right-end. Then each cytoband S, [p, q]s1s2.s3s4s5, is mapped to an integer interval I'. The left-end of S is mapped to I'start and the right-end of S maps I'stop, where I'start and I'stop are integers. That is, each cytoband becomes an integer interval I' as follows
 |
|
| (1) |
|
| (2) |
k, k' 5, sk 
0, s(k+1) = 0, tk' 0, and s(k+1) = 0.
The concept and some examples are depicted in Fig. 3. The regions 1p1 and 3q11–3q24 map to [–199999, –100000] and [110000, 249999], respectively.
 |
 |
 |
 |
|
Given two cytoband regions S = s[p, q]s1s2. s3s4s5 and T = t[p, q]t1t2.t3t4t5, we say S = T if and only if s = t and sstart = tstart, sstop = tstop (f(s1s2.s3s4s5) = f(t1t2.t3t4t5), g(s1s2.s3s4s5) = g(t1t2.t3t4t5)). In the following we will show that the mapping function can resolve the problem of CQ. First we will prove that the mapping function is one-to-one. This promises that every cytoband region can map to one integer interval and just one integer interval. Secondly, we will prove that the mapping function is monotone. This demonstrates that the order of two cytoband regions will not change during the mapping.
THEOREM 1
Functions f and g are one-to-one.
PROOF
Given two distinct cytoband regions S and T in the same chromosome, as S is not equal to T, the pair (left-end of S, right-end of S) is not equal to the pair (left-end of T, right-end of T). Let the left-end of S be the left-end of [p, q]s1s2.s3s4s5 and the left-end of T be the left-end of [p, q]t1t2.t3t4t5. Since [p, q]s1s2.s3s4s5 is not equal to [p, q]t1t2.t3t4t5, ([p, q], s1, s2, s3, s4, s5) must not be equal to ([p, q], t1, t2, t3, t4, t5). If the left-end of S is qter, cen or pter, then the left-end of T is not the same as the left-end of S. If the left-end of S and the left-end of T are in q-arm then:
where k = 1, 2, 3, 4, or 5, sk
Similarly, we can also prove that g is a one-to-one function. As both f and g are one-to-one functions, we know that the mapping function to transform a cytoband region to an integer interval is one-to-one.
THEOREM
Functions f and g are monotone.
PROOF
Let S and T be two cytoband regions. If the left-end of S is before the left-end of T, by the definition of function f in Equation (1) f(S) is smaller than f(S). If the left-end of S is equal to the left-end of S, then f(T) must be equal to f(S). If the left-end of S is behind the left-end of T, then f(S) must be greater than f(T). Hence, we conclude that function f is monotone.
In the same way, we can prove that function g is also monotone.
According to Theorems 1 and 2, we know that the mapping function is one-to-one and monotone. Therefore, the problem of CQ can be safely transformed to the problem of finding the intersections of integer intervals.
|
3 IMPLEMENTATION |
|---|
|
TOPAbstract1 INTRODUCTION2 SYSTEM AND METHODS3 IMPLEMENTATION4 PERFORMANCE EVALUATION5 DISCUSSIONS AND CONCLUSIONSREFERENCES |
|---|
The CQS has a three-tier structure, i.e. Database Server, Web Server and Clients. We implement CQS on both Windows and Linux operation systems. CQS is a web based query system running on Apache Web Server 2.0.43 with PHP programs. The data used by CQS is stored in MySQL 3.23.54, which contains cytogenetic annotations and gene information from NCBI LocusLink and NCBI Map Viewer. The Web Server is the bridge between Database Server and Clients. Users may use a browser to access CQS via Internet.
There are two stages for the construction of CQS: data preprocessing stage and query stage. In data preprocessing stage, data from NCBI LocusLink and NCBI Map Viewer is parsed and saved to a local database. During the procedure, every cytogenetic annotation is transformed to three values, say chromosome value, start value and stop value. In the first part of the query stage, every query region is also transformed to the three values as in the data preprocessing stage. In the second part of the query stage, all the genes relevant to the query regions will be retrieved according to the rules described in the following. Suppose all the cytogenetic data is stored in a database table named ‘Cytogenetic’, and the corresponding fields are ‘chromosome’, ‘start’ and ‘stop’. Let the chromosome value, start value, and stop value of query regions be expressed as ‘qchromosome’, ‘qstart’ and ‘qstop’, respectively. Then the group 1 set to group 4 set in Fig. 1 will be obtained by the four queries listed below. The complete answers are the combination of these four query results.
Query 1. SELECT * from Cytogenetic where chromosome = qchromosome and ((start < qstart) and (stop >= qstart) and (stop < qstop)).
Query 2. SELECT * from Cytogenetic where chromosome = qchromosome and ((start >= qstart) and (stop <= qstop)).
Query 3. SELECT * from Cytogenetic where chromosome = qchromosome and (((start < qstart) and (stop >= qstop)) or((start <= qstart) and (stop > qstop))).
Query 4. SELECT * from Cytogenetic where chromosome = qchromosome and ((start > qstart) and (start <= qstop) and (stop > qstop)).
|
4 PERFORMANCE EVALUATION |
|---|
|
TOPAbstract1 INTRODUCTION2 SYSTEM AND METHODS3 IMPLEMENTATION4 PERFORMANCE EVALUATION5 DISCUSSIONS AND CONCLUSIONSREFERENCES |
|---|
In order to verify the accuracy of CQS in real-world applications, the performance of CQS is tested using cytogenetic annotations from NCBI Map Viewer, which offers an interface to query genes in cytobands. If CQS performed as predicted in Section 2, all genes that are found correctly by NCBI Map Viewer should also be found to be relevant to the query regions by CQS.
NCBI Map Viewer integrates map and sequence data from a variety of sources. It provides query interfaces with diverse options for visualization. The Gene-on-Sequence interface aligns genes according to their sequence positions. The Gene-on-Cytogenetics interface maps genes to their cytobands. We parse the raw data from NCBI Map Viewer Build 34.3 and construct a local database for CQS. This ensures that CQS and NCBI Map Viewer are using the same set of cytogenetic annotations.
We generate two groups of queries for testing. One group simulates the searching of genes within one cytoband (Table 1). The other group simulates the situations when the query covers a range of cytobands (Table 2). Each group contains 24 queries generated randomly from each of the 24 different human chromosomes. The results are summarized in Tables 1 and 2. The column ‘cytoband’ refers to the query regions for testing. The column ‘Map Viewer’ refers to the number of genes resulting from queries in Gene-on-Cytogenetics. The column ‘CQS’ refers to the number of genes found to be relevant to the query regions by CQS. All query results are reviewed by two different persons and are classified as ‘Correct’ if they belong to groups 1–4 in Fig. 1. Otherwise the results are regarded as errors. The performance test shows that all correct genes found by NCBI Map Viewer are also found to be relevant by CQS, and all genes found to be relevant by CQS are correct.
|
|
In Tables 1 and 2, the column ‘Intersection’ refers to the number of genes found by both the NCBI Map Viewer and CQS. The column ‘Error number’ denotes the amount of errors resulting from the NCBI Map Viewer. Since CQS finds all correct results, ‘Error number’ equals to ‘Map Viewer’ minus ‘Intersection’. ‘Error ratio’ is the value of ‘Error number’ divided by ‘Map Viewer’. The column ‘Missing number’ refers to the amount of correct results that are not found by the NCBI Map Viewer. This equals the value of ‘CQS’ minus that of ‘Intersection’. ‘Missing ratio’ is the value of ‘Missing number’ divided by ‘CQS’. For example, when the query region is ‘1p34.1’, there are 221 results in the NCBI Map Viewer and 262 in CQS. In these two sets of query results, 194 are the same. These results are all relevant to the query region. There are 27 results in the NCBI Map Viewer which are not in CQS. These results are found to be errors, such as gene HCRTR1 in region 1p34.3 and gene ARTN in region 1p33–p32. There are 68 results in CQS which are not in the NCBI Map Viewer. They are found to be correct, such as gene BMP8B in region 1p35–p32 and gene EDN2 in region 1p34. Table 1 demonstrates that the values are 21% error rate and 23% missing rate in Gene-on-Cytogenetics when the query is in a single cytoband. Table 1 demonstrates that the values are 13% error rate and 9% missing rate in Gene-on-Cytogenetics when the query is in a range of cytobands. The complete list of test results can be reviewed in the Supplementary material.
|
5 DISCUSSIONS AND CONCLUSIONS |
|---|
|
TOPAbstract1 INTRODUCTION2 SYSTEM AND METHODS3 IMPLEMENTATION4 PERFORMANCE EVALUATION5 DISCUSSIONS AND CONCLUSIONSREFERENCES |
|---|
CQS uses a simple robust mapping function to solve the problem of CQ. It is the first and to our knowledge, the only comprehensive query system based on cytoband annotation alone. It is scalable and can be applied to any nomenclature system with the same hierarchical structure as the one used in human cytobands. For example, we extracted the rat and cow cytoband information from LocusLink and established separate CQS systems for these two species (http://moris.csie.ncku.edu.tw/cqs/).
Due to the heterogeneous quality of cytoband annotations in the current database, some genes are assigned to broad cytogenetic regions, such as 1, 1p or 1q. These genes will be included in the CQS results if their annotations encompass the query regions (group 3 in Fig. 1). The group 2 results (annotations residing in the query regions) will contain genes with more precise annotations. With the method used in CQS, the database with accurate and precise cytogenetic annotations would give rise to accurate and precise results. We hope that in the near future there will be databases with precise cytogenetic annotations for all genes.
It is worth mentioning that the problem of CQ can also be solved by sequence mapping. Such a solution requires the knowledge of the entire sequence of human genome, i.e. all DNA sequences of human chromosomes, and more importantly, it demands efforts to define the position of every gene on the sequence map. Since the sequence position is numerical, after mapping to the genome sequence, every gene becomes a numerical segment with a start position and a stop position on a numerical map. If cytobands can be properly translated into segments on the sequence map, the problem of CQ can be resolved by simple numerical calculations and set operations just like what we did in CQS. However to our knowledge, there is no comprehensive method to define the boundaries of cytobands on the sequence map. The boundaries seem to be determined empirically. As a result, genes close to the current boundaries may not be properly assigned to their corresponding cytoband segments. If they were in the ‘wrong’ cytoband segment, a query of the ‘correct’ cytoband segment will not include these genes in the final results.
There are many discrepancies between the order of sequence positions and that of the cytogenetic locations (Korenberg et al., 1999). Since technologies using metaphase chromosomes have been around for a long time, most clinical observations were expressed as cytobands in the literature. Biomedical researchers would prefer to collect all genes whose cytoband annotations reside in the clinically important cytobands, regardless of the discrepancies in the sequence mapping. However, we believe the ultimate CQS would be a combination of CQS and sequence mapping. Such a system is being developed in our laboratory.
CQS is tolerant to the most common human mistake in CQ: typing errors resulting in numbers not matching any cytogenetic annotations in the database. For example, one may submit a query ‘8q22–8q23.4’ in which cytoband 8q23.4 does not exist (Fig. 2). There are several methods to deal with this problem: (1) to design a meticulous drop-down menu that incorporates all possible cytobands for users to choose from without any typing of queries, (2) to look up a table containing all possible cytobands and send an error message to the user when the submitted query does not match any record in the table, or (3) to change the wrong cytoband into the correct one automatically (in this example, from 8q23.4 to 8q23.3) and proceed with the query process. All these methods require additional programming that may introduce bugs into the system. In CQS, the wrong query ‘8q22–8q23.4’ will produce exactly the same result as the right one ‘8q22–8q23.3’. The non-existing cytobands will not affect the performance of CQS. Consequently, no additional programming is necessary to manage this kind of typing error in CQS.
The algorithm used by CQS can also be applied to other cytogenetic problems, such as to transform the ISCN karyotypes into computer readable cytogenetic notations (Bradtke et al., 2003).
In conclusion, we have developed a simple robust CQ method, which resolves the problem incurred by the cytoband nomenclature system. Its implementation has given rise to a tool that should be found to be very useful to the biomedical researches.
|
Acknowledgments |
|---|
Grant funding for this work was from NSC 92-2320-B-006-070 to C.-L.H. and from NSC91-2321-B006-003 and NSC93-2321-B006-009 to H.-S.L. The authors would like to thank co-members in the Interdisciplinary Research Group on Bladder Cancer at NCKU for the discussions and comments on our work. The members other than the authors include Dr Nan-Haw Chow (pathology), Dr Jyh-Wei Shin (parasitology), Dr Tsuey-Yu Chang (parasitology), Dr Vincent Shin-Mu Tseng (computer science), Dr Jung-Hsien Chiang (computer science) and Dr Shih-Huang Chan (statistics) from National Cheng Kung University, Dr Yow-Ling Shiue (biomedical science) from National Chung Shan University, Dr Hung Wu and Dr Ting-Feng Wu (biotechnology) from Southern Taiwan University of Technology. This Research Group is led by H.-S.L. under the support of grants NSC91-2321-B006-003 and NSC93-2321-B006-009.
Conflict of Interest: none declared.
Received on November 26, 2004; revised on June 28, 2005; accepted on June 28, 2005
|
REFERENCES |
|---|
|
TOPAbstract1 INTRODUCTION2 SYSTEM AND METHODS3 IMPLEMENTATION4 PERFORMANCE EVALUATION5 DISCUSSIONS AND CONCLUSIONSREFERENCES |
|---|
Bradtke, J., et al. (2003) Computer aided analysis of additional chromosome aberrations in Philadelphia chromosome positive acute lymphoblastic leukaemia using a simplified computer readable cytogenetic notation. BMC Bioinformatics, 28, 4.
Caspersson, T., et al. (1968) Chemical differentiation along metaphase chromosomes. Exp. Cell Res., 49, 219–222[CrossRef][Web of Science][Medline].
Caspersson, T., et al. (1971) Identification of human chromosomes in a mouse–human hybrid by fluorescence techniques. Exp. Cell Res., 65, 475–478[CrossRef][Web of Science][Medline].
Collins, F.S. (1995) Positional cloning moves from perditional to traditional. Nat. Genet., 9, 347–350[CrossRef][Web of Science][Medline].
Kallioniemi, A., et al. (1992) Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science, 258, 818–821
Kirsch, I.R., et al. (2000) A systematic, high-resolution linkage of the cytogenetic and physical maps of the human genome. Nat. Genet., 24, 339–340[CrossRef][Web of Science][Medline].
Korenberg, J.R., et al. (1999) Human genome anatomy: BACs integrating the genetic and cytogenetic maps for bridging genome and biomedicine. Genome Res., 9, 994–1001
Liyanage, M., et al. (1996) Multicolour spectral karyotyping of mouse chromosomes. Nat. Genet., 14, 312–315[CrossRef][Web of Science][Medline].
Mitelman, F. and Heim, S. (1988) Consistent involvement of only 71 of the 329 chromosomal bands of the human genome in primary neoplasia-associated rearrangements. Cancer Res., 48, 7115–7119
Schrock, E., et al. (1996) Multicolor spectral karyotyping of human chromosomes. Science, 273, 494–497[Abstract].
Shikoshi, K., et al. (1994) [Therapy-related acute non-lymphocytic leukemia (M2) with 7;11 chromosome translocation induced into complete remission by low dose cytosine arabinoside and cytarabine ocfosfate therapy]. Nippon Ronen Igakkai Zasshi, 31, 468–471[Medline].
Weinstein, I.H., et al. (1969) [Down's syndrome due to chromosome translocation]. Prensa Med Argent, 56, 1273–1274[Medline].
Zech, L., et al. (1976) Characteristic chromosomal abnormalities in biopsies and lymphoid-cell lines from patients with Burkitt and non-Burkitt lymphomas. Int. J. Cancer, 17, 47–56[Web of Science][Medline].
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K.-H. Yen, C.-L. Ho, and C. Lee The analysis of inconsistencies between cytogenetic annotations and sequence mapping by defining the imprecision zones of cytogenetic banding Bioinformatics, April 1, 2009; 25(7): 845 - 852. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||