John C. Castle
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
John C. Castle has a total of 45 co-authors. Here are the top co-authors by number of shared publications:
- Ugur Sahin
BioNTech (Germany) 4 Mainz
5 shared publications - Sebastian Boegel
Division of Rheumatology and Clinical Immunology, Department of Internal Medicine I, University Medical Center of the Johannes Gutenberg University, 55131 Mainz, Germany
4 shared publications - Martin Löwer
TRON—Translational Oncology at the Medical Center of the Johannes Gutenberg-University Mainz Gemeinnützige GmbH, 55131 Mainz, Germany
3 shared publications - Thomas Bukur
Translational Oncology at the University Medical Center of the Johannes Gutenberg University Mainz (TRON), Germany (T.B., V.B.).
3 shared publications - Bernhard Y. Renard
Hasso Plattner Institute, Digital Engineering Faculty, University of Potsdam , Potsdam
2 shared publications - Jos de Graaf
Translational Oncology at Johannes Gutenberg-University Medical Center gGmbH, D-55131 Mainz, Germany
2 shared publications
Latest Publications
- Open Access
- 22 April 2014
- Published by Springer Nature in Scientific Reports
- Vol. 4 (1) , 4743
Mutated tumor alleles are expressed according to their DNA frequency
The transcription of tumor mutations from DNA into RNA has implications for biology, epigenetics and clinical practice. It is not clear if mutations are in general transcribed and, if so, at what proportion to the wild-type allele. Here, we examined the correlation between DNA mutation allele frequency and RNA mutation allele frequency. We sequenced the exome and transcriptome of tumor cell lines with large copy number variations, identified heterozygous single nucleotide mutations and absolute DNA copy number and determined the corresponding DNA and RNA mutation allele fraction. We found that 99% of the DNA mutations in expressed genes are expressed as RNA. Moreover, we found a high correlation between the DNA and RNA mutation allele frequency. Exceptions are mutations that cause premature termination codons and therefore activate nonsense-mediated decay. Beyond this, we did not find evidence of any wide-scale mechanism, such as allele-specific epigenetic silencing, preferentially promoting mutated or wild-type alleles. In conclusion, our data strongly suggest that genes are equally transcribed from all alleles, mutated and wild-type and thus transcribed in proportion to their DNA allele frequency.
- 22 April 2014
- Published by Springer Nature in Scientific Reports
- Vol. 4 (1) , 4743
Open AccessMutated tumor alleles are expressed according to their DNA frequency
The transcription of tumor mutations from DNA into RNA has implications for biology, epigenetics and clinical practice. It is not clear if mutations are in general transcribed and, if so, at what proportion to the wild-type allele. Here, we examined the correlation between DNA mutation allele frequency and RNA mutation allele frequency. We sequenced the exome and transcriptome of tumor cell lines with large copy number variations, identified heterozygous single nucleotide mutations and absolute DNA copy number and determined the corresponding DNA and RNA mutation allele fraction. We found that 99% of the DNA mutations in expressed genes are expressed as RNA. Moreover, we found a high correlation between the DNA and RNA mutation allele frequency. Exceptions are mutations that cause premature termination codons and therefore activate nonsense-mediated decay. Beyond this, we did not find evidence of any wide-scale mechanism, such as allele-specific epigenetic silencing, preferentially promoting mutated or wild-type alleles. In conclusion, our data strongly suggest that genes are equally transcribed from all alleles, mutated and wild-type and thus transcribed in proportion to their DNA allele frequency.
- 26 March 2014
- Published by Springer Nature
- p. 229-236
Genomics Meets Cancer Immunotherapy
- Top Cited PapersOpen Access
- 13 March 2014
- Published by Springer Nature in BMC Genomics
- Vol. 15 (1) , 1-12
Immunomic, genomic and transcriptomic characterization of CT26 colorectal carcinoma
Tumor models are critical for our understanding of cancer and the development of cancer therapeutics. Here, we present an integrated map of the genome, transcriptome and immunome of an epithelial mouse tumor, the CT26 colon carcinoma cell line. We found that Kras is homozygously mutated at p.G12D, Apc and Tp53 are not mutated, and Cdkn2a is homozygously deleted. Proliferation and stem-cell markers, including Top2a, Birc5 (Survivin), Cldn6 and Mki67, are highly expressed while differentiation and top-crypt markers Muc2, Ms4a8a (MS4A8B) and Epcam are not. Myc, Trp53 (tp53), Mdm2, Hif1a, and Nras are highly expressed while Egfr and Flt1 are not. MHC class I but not MHC class II is expressed. Several known cancer-testis antigens are expressed, including Atad2, Cep55, and Pbk. The highest expressed gene is a mutated form of the mouse tumor antigen gp70. Of the 1,688 non-synonymous point variations, 154 are both in expressed genes and in peptides predicted to bind MHC and thus potential targets for immunotherapy development. Based on its molecular signature, we predicted that CT26 is refractory to anti-EGFR mAbs and sensitive to MEK and MET inhibitors, as have been previously reported. CT26 cells share molecular features with aggressive, undifferentiated, refractory human colorectal carcinoma cells. As CT26 is one of the most extensively used syngeneic mouse tumor models, our data provide a map for the rationale design of mode-of-action studies for pre-clinical evaluation of targeted- and immunotherapies.
- 13 March 2014
- Published by Springer Nature in BMC Genomics
- Vol. 15 (1) , 1-12
Top Cited PapersOpen AccessImmunomic, genomic and transcriptomic characterization of CT26 colorectal carcinoma
Tumor models are critical for our understanding of cancer and the development of cancer therapeutics. Here, we present an integrated map of the genome, transcriptome and immunome of an epithelial mouse tumor, the CT26 colon carcinoma cell line. We found that Kras is homozygously mutated at p.G12D, Apc and Tp53 are not mutated, and Cdkn2a is homozygously deleted. Proliferation and stem-cell markers, including Top2a, Birc5 (Survivin), Cldn6 and Mki67, are highly expressed while differentiation and top-crypt markers Muc2, Ms4a8a (MS4A8B) and Epcam are not. Myc, Trp53 (tp53), Mdm2, Hif1a, and Nras are highly expressed while Egfr and Flt1 are not. MHC class I but not MHC class II is expressed. Several known cancer-testis antigens are expressed, including Atad2, Cep55, and Pbk. The highest expressed gene is a mutated form of the mouse tumor antigen gp70. Of the 1,688 non-synonymous point variations, 154 are both in expressed genes and in peptides predicted to bind MHC and thus potential targets for immunotherapy development. Based on its molecular signature, we predicted that CT26 is refractory to anti-EGFR mAbs and sensitive to MEK and MET inhibitors, as have been previously reported. CT26 cells share molecular features with aggressive, undifferentiated, refractory human colorectal carcinoma cells. As CT26 is one of the most extensively used syngeneic mouse tumor models, our data provide a map for the rationale design of mode-of-action studies for pre-clinical evaluation of targeted- and immunotherapies.
- Open Access
- 18 February 2013
- Published by Springer Nature in Behavioral and Brain Functions
- Vol. 9 (1) , 7
A boy with homozygous microdeletion of NEUROG1 presents with a congenital cranial dysinnervation disorder [Moebius syndrome variant]
Background: We report on a 6-year-old Turkish boy with profound sensorineural deafness, balance disorder, severe disorder of oral motor function, and mild developmental delay. Further findings included scaphocephaly, plagiocephaly, long palpebral fissures, high narrow palate, low-set posteriorly rotated ears, torticollis, hypoplastic genitalia and faulty foot posture. Parents were consanguineous. Methods and results: Computed tomography and magnetic resonance imaging showed bilateral single widened cochlear turn, narrowing of the internal auditory canal, and bilateral truncation of the vestibulo-cochlear nerve. Microarray analysis and next generation sequencing showed a homozygous deletion of chromosome 5q31.1 spanning 115.3 kb and including three genes: NEUROG1 (encoding neurogenin 1), DCNP1 (dendritic cell nuclear protein 1, C5ORF20) and TIFAB (TIFA-related protein). The inability to chew and swallow, deafness and balance disorder represented congenital palsies of cranial nerves V (trigeminal nerve) and VIII (vestibulo-cochlear nerve) and thus a congenital cranial dysinnervation disorder. Conclusions: Based on reported phenotypes of neurog1 null mutant mice and other vertebrates, we strongly propose NEUROG1 as the causative gene in this boy. The human NEUROG1 resides within the DFNB60 locus for non-syndromic autosomal recessive deafness on chromosome 5q22-q31, but linkage data have excluded it from being causative in the DFNB60 patients. Given its large size (35 Mb, >100 genes), the 5q22-q31 area could harbor more than one deafness gene. We propose NEUROG1 as a new gene for syndromic autosomal recessive hearing loss and congenital cranial dysinnervation disorder including cranial nerves V and VIII.
- 18 February 2013
- Published by Springer Nature in Behavioral and Brain Functions
- Vol. 9 (1) , 7
Open AccessA boy with homozygous microdeletion of NEUROG1 presents with a congenital cranial dysinnervation disorder [Moebius syndrome variant]
Background: We report on a 6-year-old Turkish boy with profound sensorineural deafness, balance disorder, severe disorder of oral motor function, and mild developmental delay. Further findings included scaphocephaly, plagiocephaly, long palpebral fissures, high narrow palate, low-set posteriorly rotated ears, torticollis, hypoplastic genitalia and faulty foot posture. Parents were consanguineous. Methods and results: Computed tomography and magnetic resonance imaging showed bilateral single widened cochlear turn, narrowing of the internal auditory canal, and bilateral truncation of the vestibulo-cochlear nerve. Microarray analysis and next generation sequencing showed a homozygous deletion of chromosome 5q31.1 spanning 115.3 kb and including three genes: NEUROG1 (encoding neurogenin 1), DCNP1 (dendritic cell nuclear protein 1, C5ORF20) and TIFAB (TIFA-related protein). The inability to chew and swallow, deafness and balance disorder represented congenital palsies of cranial nerves V (trigeminal nerve) and VIII (vestibulo-cochlear nerve) and thus a congenital cranial dysinnervation disorder. Conclusions: Based on reported phenotypes of neurog1 null mutant mice and other vertebrates, we strongly propose NEUROG1 as the causative gene in this boy. The human NEUROG1 resides within the DFNB60 locus for non-syndromic autosomal recessive deafness on chromosome 5q22-q31, but linkage data have excluded it from being causative in the DFNB60 patients. Given its large size (35 Mb, >100 genes), the 5q22-q31 area could harbor more than one deafness gene. We propose NEUROG1 as a new gene for syndromic autosomal recessive hearing loss and congenital cranial dysinnervation disorder including cranial nerves V and VIII.
- Open Access
- 22 December 2012
- Published by Springer Nature in Genome Medicine
- Vol. 4 (12) , 102
HLA typing from RNA-Seq sequence reads
We present a method, seq2HLA, for obtaining an individual's human leukocyte antigen (HLA) class I and II type and expression using standard next generation sequencing RNA-Seq data. RNA-Seq reads are mapped against a reference database of HLA alleles, and HLA type, confidence score and locus-specific expression level are determined. We successfully applied seq2HLA to 50 individuals included in the HapMap project, yielding 100% specificity and 94% sensitivity at a P-value of 0.1 for two-digit HLA types. We determined HLA type and expression for previously un-typed Illumina Body Map tissues and a cohort of Korean patients with lung cancer. Because the algorithm uses standard RNA-Seq reads and requires no change to laboratory protocols, it can be used for both existing datasets and future studies, thus adding a new dimension for HLA typing and biomarker studies.
- 22 December 2012
- Published by Springer Nature in Genome Medicine
- Vol. 4 (12) , 102
Open AccessHLA typing from RNA-Seq sequence reads
We present a method, seq2HLA, for obtaining an individual's human leukocyte antigen (HLA) class I and II type and expression using standard next generation sequencing RNA-Seq data. RNA-Seq reads are mapped against a reference database of HLA alleles, and HLA type, confidence score and locus-specific expression level are determined. We successfully applied seq2HLA to 50 individuals included in the HapMap project, yielding 100% specificity and 94% sensitivity at a P-value of 0.1 for two-digit HLA types. We determined HLA type and expression for previously un-typed Illumina Body Map tissues and a cohort of Korean patients with lung cancer. Because the algorithm uses standard RNA-Seq reads and requires no change to laboratory protocols, it can be used for both existing datasets and future studies, thus adding a new dimension for HLA typing and biomarker studies.
- Open Access
- 27 September 2012
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 8 (9) , e1002714
Confidence-based Somatic Mutation Evaluation and Prioritization
Next generation sequencing (NGS) has enabled high throughput discovery of somatic mutations. Detection depends on experimental design, lab platforms, parameters and analysis algorithms. However, NGS-based somatic mutation detection is prone to erroneous calls, with reported validation rates near 54% and congruence between algorithms less than 50%. Here, we developed an algorithm to assign a single statistic, a false discovery rate (FDR), to each somatic mutation identified by NGS. This FDR confidence value accurately discriminates true mutations from erroneous calls. Using sequencing data generated from triplicate exome profiling of C57BL/6 mice and B16-F10 melanoma cells, we used the existing algorithms GATK, SAMtools and SomaticSNiPer to identify somatic mutations. For each identified mutation, our algorithm assigned an FDR. We selected 139 mutations for validation, including 50 somatic mutations assigned a low FDR (high confidence) and 44 mutations assigned a high FDR (low confidence). All of the high confidence somatic mutations validated (50 of 50), none of the 44 low confidence somatic mutations validated, and 15 of 45 mutations with an intermediate FDR validated. Furthermore, the assignment of a single FDR to individual mutations enables statistical comparisons of lab and computation methodologies, including ROC curves and AUC metrics. Using the HiSeq 2000, single end 50 nt reads from replicates generate the highest confidence somatic mutation call set. Next generation sequencing (NGS) has enabled unbiased, high throughput discovery of genetic variations and somatic mutations. However, the NGS platform is still prone to errors resulting in inaccurate mutation calls. A statistical measure of the confidence of putative mutation calls would enable researchers to prioritize and select mutations in a robust manner. Here we present our development of a confidence score for mutations calls and apply the method to the identification of somatic mutations in B16 melanoma. We use NGS exome resequencing to profile triplicates of both the reference C57BL/6 mice and the B16-F10 melanoma cells. These replicate data allow us to formulate the false discovery rate of somatic mutations as a statistical quantity. Using this method, we show that 50 of 50 high confidence mutation calls are correct while 0 of 44 low confidence mutations are correct, demonstrating that the method is able to correctly rank mutation calls.
- 27 September 2012
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 8 (9) , e1002714
Open AccessConfidence-based Somatic Mutation Evaluation and Prioritization
Next generation sequencing (NGS) has enabled high throughput discovery of somatic mutations. Detection depends on experimental design, lab platforms, parameters and analysis algorithms. However, NGS-based somatic mutation detection is prone to erroneous calls, with reported validation rates near 54% and congruence between algorithms less than 50%. Here, we developed an algorithm to assign a single statistic, a false discovery rate (FDR), to each somatic mutation identified by NGS. This FDR confidence value accurately discriminates true mutations from erroneous calls. Using sequencing data generated from triplicate exome profiling of C57BL/6 mice and B16-F10 melanoma cells, we used the existing algorithms GATK, SAMtools and SomaticSNiPer to identify somatic mutations. For each identified mutation, our algorithm assigned an FDR. We selected 139 mutations for validation, including 50 somatic mutations assigned a low FDR (high confidence) and 44 mutations assigned a high FDR (low confidence). All of the high confidence somatic mutations validated (50 of 50), none of the 44 low confidence somatic mutations validated, and 15 of 45 mutations with an intermediate FDR validated. Furthermore, the assignment of a single FDR to individual mutations enables statistical comparisons of lab and computation methodologies, including ROC curves and AUC metrics. Using the HiSeq 2000, single end 50 nt reads from replicates generate the highest confidence somatic mutation call set. Next generation sequencing (NGS) has enabled unbiased, high throughput discovery of genetic variations and somatic mutations. However, the NGS platform is still prone to errors resulting in inaccurate mutation calls. A statistical measure of the confidence of putative mutation calls would enable researchers to prioritize and select mutations in a robust manner. Here we present our development of a confidence score for mutations calls and apply the method to the identification of somatic mutations in B16 melanoma. We use NGS exome resequencing to profile triplicates of both the reference C57BL/6 mice and the B16-F10 melanoma cells. These replicate data allow us to formulate the false discovery rate of somatic mutations as a statistical quantity. Using this method, we show that 50 of 50 high confidence mutation calls are correct while 0 of 44 low confidence mutations are correct, demonstrating that the method is able to correctly rank mutation calls.
- Open Access
- 4 August 2011
- Published by Springer Nature in BMC Bioinformatics
- Vol. 12 (1) , 324
rapmad: Robust analysis of peptide microarray data
Peptide microarrays offer an enormous potential as a screening tool for peptidomics experiments and have recently seen an increased field of application ranging from immunological studies to systems biology. By allowing the parallel analysis of thousands of peptides in a single run they are suitable for high-throughput settings. Since data characteristics of peptide microarrays differ from DNA oligonucleotide microarrays, computational methods need to be tailored to these specifications to allow a robust and automated data analysis. While follow-up experiments can ensure the specificity of results, sensitivity cannot be recovered in later steps. Providing sensitivity is thus a primary goal of data analysis procedures. To this end we created rapmad (Robust Alignment of Peptide MicroArray Data), a novel computational tool implemented in R.
- 4 August 2011
- Published by Springer Nature in BMC Bioinformatics
- Vol. 12 (1) , 324
Open Accessrapmad: Robust analysis of peptide microarray data
Peptide microarrays offer an enormous potential as a screening tool for peptidomics experiments and have recently seen an increased field of application ranging from immunological studies to systems biology. By allowing the parallel analysis of thousands of peptides in a single run they are suitable for high-throughput settings. Since data characteristics of peptide microarrays differ from DNA oligonucleotide microarrays, computational methods need to be tailored to these specifications to allow a robust and automated data analysis. While follow-up experiments can ensure the specificity of results, sensitivity cannot be recovered in later steps. Providing sensitivity is thus a primary goal of data analysis procedures. To this end we created rapmad (Robust Alignment of Peptide MicroArray Data), a novel computational tool implemented in R.
- Open Access
- 6 June 2011
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 6 (6) , e20660
SNPs Occur in Regions with Less Genomic Sequence Conservation
Rates of SNPs (single nucleotide polymorphisms) and cross-species genomic sequence conservation reflect intra- and inter-species variation, respectively. Here, I report SNP rates and genomic sequence conservation adjacent to mRNA processing regions and show that, as expected, more SNPs occur in less conserved regions and that functional regions have fewer SNPs. Results are confirmed using both mouse and human data. Regions include protein start codons, 3′ splice sites, 5′ splice sites, protein stop codons, predicted miRNA binding sites, and polyadenylation sites. Throughout, SNP rates are lower and conservation is higher at regulatory sites. Within coding regions, SNP rates are highest and conservation is lowest at codon position three and the fewest SNPs are found at codon position two, reflecting codon degeneracy for amino acid encoding. Exon splice sites show high conservation and very low SNP rates, reflecting both splicing signals and protein coding. Relaxed constraint on the codon third position is dramatically seen when separating exonic SNP rates based on intron phase. At polyadenylation sites, a peak of conservation and low SNP rate occurs from 30 to 17 nt preceding the site. This region is highly enriched for the sequence AAUAAA, reflecting the location of the conserved polyA signal. miRNA 3′ UTR target sites are predicted incorporating interspecies genomic sequence conservation; SNP rates are low in these sites, again showing fewer SNPs in conserved regions. Together, these results confirm that SNPs, reflecting recent genetic variation, occur more frequently in regions with less evolutionarily conservation.
- 6 June 2011
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 6 (6) , e20660
Open AccessSNPs Occur in Regions with Less Genomic Sequence Conservation
Rates of SNPs (single nucleotide polymorphisms) and cross-species genomic sequence conservation reflect intra- and inter-species variation, respectively. Here, I report SNP rates and genomic sequence conservation adjacent to mRNA processing regions and show that, as expected, more SNPs occur in less conserved regions and that functional regions have fewer SNPs. Results are confirmed using both mouse and human data. Regions include protein start codons, 3′ splice sites, 5′ splice sites, protein stop codons, predicted miRNA binding sites, and polyadenylation sites. Throughout, SNP rates are lower and conservation is higher at regulatory sites. Within coding regions, SNP rates are highest and conservation is lowest at codon position three and the fewest SNPs are found at codon position two, reflecting codon degeneracy for amino acid encoding. Exon splice sites show high conservation and very low SNP rates, reflecting both splicing signals and protein coding. Relaxed constraint on the codon third position is dramatically seen when separating exonic SNP rates based on intron phase. At polyadenylation sites, a peak of conservation and low SNP rate occurs from 30 to 17 nt preceding the site. This region is highly enriched for the sequence AAUAAA, reflecting the location of the conserved polyA signal. miRNA 3′ UTR target sites are predicted incorporating interspecies genomic sequence conservation; SNP rates are low in these sites, again showing fewer SNPs in conserved regions. Together, these results confirm that SNPs, reflecting recent genetic variation, occur more frequently in regions with less evolutionarily conservation.