A MULTILEVEL MODEL TO ADDRESS BATCH EFFECTS IN COPY NUMBER ESTIMATION USING SNP ARRAYS

Submicroscopic changes in chromosomal DNA copy number dosage are common and have been implicated in many heritable diseases and cancers. Recent high-throughput technologies have a resolution that permits the detection of segmental changes in DNA copy number that span thousands of basepairs across the genome. Genome-wide association studies (GWAS) may simultaneously screen for copy number-phenotype and SNP-phenotype associations as part of the analytic strategy. However, genome-wide array analyses are particularly susceptible to batch effects as the logistics of preparing DNA and processing thousands of arrays often involves multiple laboratories and technicians, or changes over calendar time to the reagents and laboratory equipment. Failure to adjust for batch effects can lead to incorrect inference and requires inefficient post-hoc quality control procedures that exclude regions that are associated with batch. Our work extends previous model-based approaches for copy number estimation by explicitly modeling batch effects and using shrinkage to improve locus-specific estimates of copy number uncertainty. Key features of this approach include the use of diallelic genotype calls from experimental data to estimate batch- and locus-specific parameters of background and signal without the requirement of training data. We illustrate these ideas using a study of bipolar disease and a study of chromosome 21 trisomy. The former has batch effects that dominate much of the observed variation in quantile-normalized intensities, while the latter illustrates the robustness of our approach to datasets where as many as 25% of the samples have altered copy number. Locus-specific estimates of copy number can be plotted on the copy-number scale to investigate mosaicism and guide the choice of appropriate downstream approaches for smoothing the copy number as a function of physical position. The software is open source and implemented in the R package CRLMM available at Bioconductor (http:www.bioconductor.org).

Polymorphism and chromosomal location of the MC4R (melanocortin-4 receptor) gene in the dog and red fox

The melanocortin-4 receptor (MC4R) is expressed in the hypothalamus and regulates energy intake and body weight. In silico screening of the canine chromosome 1 sequence and a comparison with the porcine MC4R sequence by BLAST were performed. The nucleotide sequence of the whole coding region and 3'- and 5'-flanking regions of the dog (1214 bp) and red fox (1177 bp) MC4R gene was established and high conservation of the nucleotide sequences was revealed (99%). Five sets of PCR primers were designed and a search for polymorphism was performed by the SSCP technique in a group of 31 dogs representing nineteen breeds and 35 farm red foxes. Sequencing of DNA fragments, representing the identified SSCP patterns, revealed three single nucleotide polymorphisms (including a missense one) in dogs and four silent SNPs in red foxes. An average SNP frequency was approx. 1/400 bp in the dog and 1/300 bp in the red fox. We mapped the MC4R gene by FISH to the canine chromosome 1 (CFA1q1.1) and to the red fox chromosome 5 (VVU5p1.2).

Losses of chromosome arms 4q, 8p, 13q and gain of 8q are correlated with increasing chromosomal instability in hepatocellular carcinoma

OBJECTIVE: Chromosomal instability is a key feature in hepatocellular carcinoma (HCC). Array comparative genomic hybridization (aCGH) revealed recurring structural aberrations, whereas fluorescence in situ hybridization (FISH) indicated an increasing number of numerical aberrations in dedifferentiating HCC. Therefore, we examined whether there was a correlation between structural and numerical aberrations of chromosomal instability in HCC. METHODS AND RESULTS: 27 HCC (5 well, 10 moderately, 12 lower differentiated) already cytogenetically characterized by aCGH were analyzed. FISH analysis using probes for chromosomes 1, 3, 7, 8 and 17 revealed 1.46-4.24 signals/nucleus, which correlated with the histological grade (well vs. moderately,p < 0.0003; moderately vs. lower, p < 0.004). The number of chromosomes to each other was stable with exceptions only seen for chromosome 8. Loss of 4q and 13q, respectively, were correlated with the number of aberrations detected by aCGH (p < 0.001, p < 0.005; Mann-Whitney test). Loss of 4q and gain of 8q were correlated with an increasing number of numerical aberrations detected by FISH (p < 0.020, p < 0.031). Loss of 8p was correlated with the number of structural imbalances seen in aCGH (p < 0.048), but not with the number of numerical changes seen in FISH. CONCLUSION: We found that losses of 4q, 8p and 13q were closely correlated with an increasing number of aberrations detected by aCGH, whereas a loss of 4q and a gain of 8q were also observed in the context of polyploidization, the cytogenetic correlate of morphological dedifferentiation.

A chromosomal region on ECA13 is associated with maxillary prognathism in horses.

Hereditary variations in head morphology and head malformations are known in many species. The most common variation encountered in horses is maxillary prognathism. Prognathism and brachygnathism are syndromes of the upper and lower jaw, respectively. The resulting malocclusion can negatively affect teeth wear, and is considered a non-desirable trait in breeding programs. We performed a case-control analysis for maxillary prognathism in horses using 96 cases and 763 controls. All horses had been previously genotyped with a commercially available 50 k SNP array. We analyzed the data with a mixed-model considering the genomic relationships in order to account for population stratification. Two SNPs within a region on the distal end of chromosome ECA 13 reached the Bonferroni corrected genome-wide significance level. There is no known prognathism candidate gene located within this region. Therefore, our findings in the horse offer the possibility of identifying a novel gene involved in the complex genetics of prognathism that might also be relevant for humans and other livestock species.

Chromosomal control of the dominant suppression of {\it ras\/}-mediated transformation

The role of tumor suppressor function in the multistep process of carcinogenesis was studied in the human teratocarcinoma cell line PA-1. Early passage PA-1 cells ($<$P100) are preneoplastic while late passage ($>$P100) PA-1 cells are spontaneously transformed. Previous work demonstrated a causal role for the N-ras oncogene in the neoplastic transformation of this cell line and the gene was cloned. A clonal cell line established at passage 40 has been shown to suppress the neoplastic transformation potential of the PA-1 N-ras oncogene in gene transfer experiments. This phenotype has been termed SRT+ for suppression of ras transformation. A clonal cell line established at passage 63 is neoplastically transformed by the N-ras in similar gene transfer experiments and is regarded as srt$-$. Somatic cell hybrids were formed between the SRT+ cell and two different N-ras transformed srt$-$ cells. The results indicate that five of the seven independent hybrid clones, and all 14 subclones, failed to form tumors in the nude mouse tumor assay. Chromosomal analysis of rare neoplastic segregants which arose from suppressed hybrid populations demonstrate that the general loss of chromosomes correlates with the reemergence of neoplastic transformation. Karyotype analyses demonstrate a statistically correlative loss of chromosomes 1, 4, 19, and to a lesser extent 11, 14, and 16. DNA hybridization analysis demonstrates a single copy of the intact N-ras oncogene in parental cells, suppressed hybrids, and neoplastically transformed hybrids. These results indicate that functional ras transformation suppression is a trans-dominant trait which may be controlled by sequences residing on particular chromosomes in the human genome. Furthermore, the suppression of ras transformation results from a unique step in the multistep process of carcinogenesis that is different from the induction of immortality. Thus, the neoplastic process of the PA-1 cell line involves at least three steps: (1) induction of immortality, (2) activation of the N-ras oncogene, and (3) loss of tumor suppressor function. ^

Chromosomal organization and evolutionary conservation of the human chromosome 19q linkage group containing three DNA repair genes

A series of human-rodent somatic cell hybrids were investigated by Southern blot analysis for the presence or absence of twenty-six molecular markers and three isozyme loci from human chromosome 19. Based on the co-retention of these markers in the various independent hybrid clones containing portions of human chromosome 19 and on pulsed field mapping, chromosome 19 is divided into twenty ordered regions. The most likely marker order for the chromosome is: (LDLR, C3)-(cen-MANNB)-D19S7-PEPD-D19S9-GPI-TGF$ \beta$-(CYP2A, NCA, CGM2, BCKAD)-PSG1a-(D19S8, XRCC1)-(D19S19, ATP1A3)-(D19S37, APOC2)-CKMM-ERCC2-ERCC1-(D19S62, D19S51)-D19S6-D19S50-D19S22-(CGB, FTL)-qter.^ The region of 19q between the proximal marker D19S7 and the distal gene coding for the beta subunit of chorionic gonadotropin (CGB) is about 37 Mb in size and covers about 37 cM genetic distance. The ration of genetic to physical distance on 19q is therefore very close to the genomic average OF 1 cM/Mb. Estimates of physical distances for intervals between chromosome 19 markers were calculated using a mapping function which estimates distances based on the number of breaks in hybrid clone panels. The consensus genetic distances between individual markers (established at HBM10) were compared to these estimates of physical distances. The close agreement between the two estimates suggested that spontaneously broken hybrids are as appropriate for this type of study as radiation hybrids.^ All three DNA repair genes located on chromosome 19 were found to have homologues on Chinese hamster chromosome 9, which is hemizygous in CHO cells, providing an explanation for the apparent ease with which mutations at these loci were identified in CHO cells. Homologues of CKMM and TGF$\beta$ (from human chromosome 19q) and a mini-satellite DNA specific to the distal region of human chromosome 19q were also mapped to Chinese hamster 9. Markers from 19p did not map to this hamster chromosome. Thus the q-arm of chromosome 19, at least between the genes PEPD and ERCC1, appears to be a linkage group which is conserved intact between humans and Chinese hamsters. ^

Control of chromosomal rearrangements in {\it Escherichia coli\/}

A complete physical map of Escherichia coli K-12 strain MG1655 was constructed by digesting chromosomal DNA with the infrequently cutting restriction enzymes NotI, SfiI and XbaI and separating the fragments by pulsed field gel electrophoresis. The map was used to compare six K-12 strains of E. coli. Although several differences were noted and localized, the map of MG1655 was representative of all the K-12 strains tested. The maps were also used to analyze chromosomal rearrangements in the E. coli strain MG1655. The spontaneous and UV induced frequencies of tandem duplication formation were measured at several loci distributed around the chromosome. The spontaneous duplication frequency varied from 10$\sp{-5}$ to 10$\sp{-3}$ and increased at least ten-fold following mild UV irradiation treatment. Duplications of several regions of the chromosome, including the serA region and the metE region, were mapped using pulsed field gel electrophoresis. Duplications of serA were found to be large, ranging in size from 600 kb to 2100 kb. Several of the duplications isolated at serA were caused by ectopic recombination between IS5 elements and between IS186 elements. Duplications of the metE region, however, were almost exclusively the result of ectopic recombination between ribosomal RNA cistrons. Duplication frequencies were determined at both serA and metE in wild type and mismatch repair mutant strains (mutL, mutS, uvrD and recF). Even though all of the mismatch repair mutations increased duplication frequency of metE, the largest increases were observed in the mutL and mutS strains. Duplication frequency of serA was increased less dramatically by mutations in mismatch repair. Several duplications of metE isolated in a wild type and a mismatch repair mutant were mapped. The results showed that the same repeated sequences were used for duplication formation in the mismatch repair mutant as were used in the wild type strain. Several isolates showed evidence of multiple rearrangements indicating that mismatch repair may play a role in stabilizing the genome by controlling chromosomal rearrangement. ^

Epidemiology of secondary leukemia with reference to chromosomal abnormalities

Secondary acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) have been recognized as one of the most feared long-term complications of cancer therapy. The aim of this case-control study was to determine the prevalence of chromosomal abnormalities and family history of cancer among secondary AML/MDS cases and de novo AML/MDS controls. Study population were 332 MD Anderson Cancer Center patients who were registered between 1986 and 1994. Cases were patients who had a prior invasive cancer before diagnoses of AML/MDS and controls were de novo AML/MDS. Cases (166) and controls (166) were frequency matched on age $\pm$5 years, sex and year of diagnosis of leukemia. Cytogenetic data were obtained from the leukemia clinic database of MD Anderson Cancer Center and data on family history of cancer and other risk factors were abstracted from the patients' medical record. The distribution of AML and MDS among cases was 58% and 42% respectively and among controls 67% and 33% respectively. Prevalence of chromosomal abnormalities were observed more frequently among cases than controls. Reporting of family history of cancer were similar among both groups. Univariate analysis revealed an odds ratio (OR) of 2.8 (95% CI 1.5-5.4) for deletion of chromosome 7, 1.9 (95% CI 0.9-3.8) for deletion of chromosome 5, 2.3 (95% CI 0.8-6.2) for deletion of 5q, 2.0 (95% CI 1.0-4.2) for trisomy 8, 1.3 (95% CI 0.8-2.1) for chromosomal abnormalities other than chromosome 5 or 7 and 1.3 (95% CI 0.8-2.0) for family history of cancer in a first degree relative. The OR remained significant for deletion of chromosome 7 (2.3, 95% CI 1.1-4.8) after adjustment for age, alcohol, smoking, occupation related to chemical exposure and family history of cancer in a first degree relative. Of the 166 secondary AML/MDS patients 70% had a prior solid tumor and 30% experienced hematological cancers. The most frequent cancers were breast (21.1%), non-Hodgkin lymphoma (13.3%), Hodgkin's disease (10.2%), prostate (7.2%), colon (6%), multiple myeloma (3.6%) and testes (3.0%). The majority of these cancer patients were treated with chemotherapy or radiotherapy or both. Abnormalities of chromosome 5 or 7 were found to be more frequent in secondary AML/MDS patients with prior hematological cancer than patients with prior solid tumors. Median time to develop secondary AML/MDS was 5 years. However, secondary AML/MDS among patients who received chemotherapy and had a family history of cancer in a first degree relative occurred earlier (median 2.25 $\pm$ 0.9 years) than among patients without such family history (median 5.50 $\pm$ 0.18 years) (p $<$.03). The implication of exposure to chemotherapy among patients with a family history of cancer needs to be further investigated. ^

CHROMOSOMAL MAPPING, LINKAGE RELATIONSHIPS, AND EVOLUTIONARY STUDIES OF THE HUMAN ANONYMOUS DNA CLONE D1S1 (IN SITU HYBRIDIZATION, PRIMATES, GENE MAPPING, AUTOSOMAL DOMINANT RETINITIS PIGMENTOSA, GENE DUPLICATION)

D1S1, an anonymous human DNA clone originally called (lamda)Ch4-H3 or (lamda)H3, was the first single copy mapped to a human chromosome (1p36) by in situ hybridization. The chromosomal assignment has been confirmed in other laboratories by repeating the in situ hybridization but not by another method. In the present study, hybridization to a panel of hamster-human somatic cell hybrids revealed copies of D1S1 on both chromosomes 1 and 3. Subcloning D1S1 showed that the D1S1 clone itself is from chromosome 3, and the sequence detected by in situ hybridization is at least two copies of part of the chromosome 3 copy. This finding demonstrates the importance of verifying gene mapping with two methods and questions the accuracy of in situ hybridization mapping.^ Non-human mammals have only one copy of D1S1, and the non-human primate D1S1 map closely resembles the human chromosome 3 copy. Thus, the human chromosome 1 copies appear to be part of a very recent duplication that occurred after the divergence between humans and the other great apes.^ A moderately informative HindIII D1S1 RFLP was mapped to chromosome 3. This marker and 12 protein markers were applied to a linkage study of autosomal dominant retinitis pigmentosa (ADRP). None of the markers proved linkage, but adding the three families examined to previously published data raises the ADRP:Rh lod score to 1.92 at (THETA) = 0.30. ^

CHROMOSOMAL ORIGIN OF DM-DNA

Double minutes (dm) are small chromatin particles of 0.3 microns diameter found only in the metaphase cells of human and murine tumors. Dm are unique cytogenetic structures since their numbers per cell show wide variation. At cell division, dm are retained despite the lack of centromeres. In squash preparations, dm show clustering often in association with chromosomes. Human carcinoma cell line SW613-S18 was found to have large numbers of dm and biological characteristics favorable for mitotic synchronization and chromosome isolation experiments.^ S18 cells were synchronized to mitosis with metabolic and mitotic blocking compounds. Mitotic cells were lysed to release chromosomes and dm from the mitotic spindle and the resulting suspensions were fractionated to enrich for dm. The DNA in enriched fractions was characterized. The reassociation kinetics of dm-DNA driven with placental human DNA was similar to the reassociation curve of labeled placental DNA under similar conditions. In situ hybridization of dm-DNA to tumor and normal metaphase cells showed grain localization over the entire karyotype. Dm-DNA was shown by pulse chase DNA replication experiments to replicate during early and mid S-phase of the cell cycle, but not in late S-phase. In addition, BrdUrd incorporation studies showed that dm-DNA replicates only once during the S-phase. Premature chromosome condensation studies suggest the basis of numerical heterogeneity of dm is nondisjunction, not anomalous or unscheduled DNA replication.^ These data and previous cytochemical banding studies of dm in SW613-S18 indicate that dm-DNA is chromosomal in origin. No evidence of gene amplification was found in the DNA reassociation data. It is likely that dm-DNA represents the pale-staining G-band regions of the human karyotype in this cell line. ^