Limits...
A method for metagenomics of Helicobacter pylori from archived formalin-fixed gastric biopsies permitting longitudinal studies of carcinogenic risk.

Zheng Z, Andersson AF, Ye W, Nyrén O, Normark S, Engstrand L - PLoS ONE (2011)

Bottom Line: About 82% and 60% of the predicted genes in the two genomes were captured by at least a single sequencing read.Along with sequences displaying high similarity to known H. pylori genes, novel and highly variant H. pylori sequences were identified in the FFPE sections by our physical enrichment approach, which would likely not have been detected by a sequence capture approach.The study demonstrates the feasibility of longitudinal metagenomic studies of H. pylori using decade-preserved FFPE biopsies.

View Article: PubMed Central - PubMed

Affiliation: Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden.

ABSTRACT
The human microbiota has come into focus in the search for component causes of chronic diseases, such as gastrointestinal cancers. Presumably long induction periods and altered local environments after disease onset call for the development of methods for characterization of microorganisms colonizing the host decades before disease onset. Sequencing of microbial genomes in old formalin-fixed and paraffin-embedded (FFPE) gastrointestinal biopsies provides a means for such studies but is still challenging. Here we report a method based on laser capture micro-dissection and modified Roche 454 high-throughput pyrosequencing to obtain metagenomic profiles of Helicobacter pylori. We applied this method to two 15 year old FFPE biopsies from two patients. Frozen homogenized biopsies from the same gastroscopy sessions were also available for comparison after re-culture of H. pylori. For both patients, H. pylori DNA dissected from FFPE sections had ~96.4% identity with culture DNA from the same patients, while only ~92.5% identity with GenBank reference genomes, and with culture DNA from the other patient. About 82% and 60% of the predicted genes in the two genomes were captured by at least a single sequencing read. Along with sequences displaying high similarity to known H. pylori genes, novel and highly variant H. pylori sequences were identified in the FFPE sections by our physical enrichment approach, which would likely not have been detected by a sequence capture approach. The study demonstrates the feasibility of longitudinal metagenomic studies of H. pylori using decade-preserved FFPE biopsies.

Show MeSH

Related in: MedlinePlus

Sequence distributions of FFPE 1 across all contigs of Culture 1 (de novo assembly) at resolutions of single-base and 1000-bp bin, and across the chromosome of strain 26695 at resolutions of single-base and open reading frame (ORF).
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3198776&req=5

pone-0026442-g004: Sequence distributions of FFPE 1 across all contigs of Culture 1 (de novo assembly) at resolutions of single-base and 1000-bp bin, and across the chromosome of strain 26695 at resolutions of single-base and open reading frame (ORF).

Mentions: Mapping of the FFPE sequences onto the assembled contigs or onto the ten reference genomes showed that the FFPE sequences were fairly evenly distributed (Figure 4 and supplementary Figures S1 and S2). The FFPE 1 sequences covered 22.1% of all Culture 1 assembled contigs at single-base resolution, and 82.2% at 1000-bp bin resolution. When mapping onto the strain 26695 genome, the corresponding percentages were 19.6% and 69.7% (Figure 4); and for strain B8 19.2% and 66.9% (plasmid 63.8% and 100%); B38 19.4% and 72.1%; G27 19.7% and 72% (plasmid 72.1% and 100%); HPAG1 19.8% and 71.8% (plasmid 72.1% and 100%); J99 18.6% and 70.2%; P12 19.2% and 69.8% (plasmid 66.4% and 90%); PeCan4 18.9% and 68.7% (plasmid 37.3% and 75%); Shi470 19.4% and 68.2%; and SJM180 18.9% and 68.3% (Supplementary Figure S1). For FFPE 2 sequences, apart from lower mapping percentages (Culture 2 12.6% and 60.1%, GenBank strains about 11% and 50%), a major difference compared with FFPE1 results was that there was nearly no sequence mapped to the plasmids (see details in Supplementary Figure S2). GC content is one of the main factors influencing amplification bias [3]. The GC content of amplified FFPE H. pylori sequences in sample FFPE 1 and FFPE 2 was 40.06% and 40.99%, respectively, close to the range of 38.89% to 39.19% of reference H. pylori genomes (UCSC genome browser).


A method for metagenomics of Helicobacter pylori from archived formalin-fixed gastric biopsies permitting longitudinal studies of carcinogenic risk.

Zheng Z, Andersson AF, Ye W, Nyrén O, Normark S, Engstrand L - PLoS ONE (2011)

Sequence distributions of FFPE 1 across all contigs of Culture 1 (de novo assembly) at resolutions of single-base and 1000-bp bin, and across the chromosome of strain 26695 at resolutions of single-base and open reading frame (ORF).
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3198776&req=5

pone-0026442-g004: Sequence distributions of FFPE 1 across all contigs of Culture 1 (de novo assembly) at resolutions of single-base and 1000-bp bin, and across the chromosome of strain 26695 at resolutions of single-base and open reading frame (ORF).
Mentions: Mapping of the FFPE sequences onto the assembled contigs or onto the ten reference genomes showed that the FFPE sequences were fairly evenly distributed (Figure 4 and supplementary Figures S1 and S2). The FFPE 1 sequences covered 22.1% of all Culture 1 assembled contigs at single-base resolution, and 82.2% at 1000-bp bin resolution. When mapping onto the strain 26695 genome, the corresponding percentages were 19.6% and 69.7% (Figure 4); and for strain B8 19.2% and 66.9% (plasmid 63.8% and 100%); B38 19.4% and 72.1%; G27 19.7% and 72% (plasmid 72.1% and 100%); HPAG1 19.8% and 71.8% (plasmid 72.1% and 100%); J99 18.6% and 70.2%; P12 19.2% and 69.8% (plasmid 66.4% and 90%); PeCan4 18.9% and 68.7% (plasmid 37.3% and 75%); Shi470 19.4% and 68.2%; and SJM180 18.9% and 68.3% (Supplementary Figure S1). For FFPE 2 sequences, apart from lower mapping percentages (Culture 2 12.6% and 60.1%, GenBank strains about 11% and 50%), a major difference compared with FFPE1 results was that there was nearly no sequence mapped to the plasmids (see details in Supplementary Figure S2). GC content is one of the main factors influencing amplification bias [3]. The GC content of amplified FFPE H. pylori sequences in sample FFPE 1 and FFPE 2 was 40.06% and 40.99%, respectively, close to the range of 38.89% to 39.19% of reference H. pylori genomes (UCSC genome browser).

Bottom Line: About 82% and 60% of the predicted genes in the two genomes were captured by at least a single sequencing read.Along with sequences displaying high similarity to known H. pylori genes, novel and highly variant H. pylori sequences were identified in the FFPE sections by our physical enrichment approach, which would likely not have been detected by a sequence capture approach.The study demonstrates the feasibility of longitudinal metagenomic studies of H. pylori using decade-preserved FFPE biopsies.

View Article: PubMed Central - PubMed

Affiliation: Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden.

ABSTRACT
The human microbiota has come into focus in the search for component causes of chronic diseases, such as gastrointestinal cancers. Presumably long induction periods and altered local environments after disease onset call for the development of methods for characterization of microorganisms colonizing the host decades before disease onset. Sequencing of microbial genomes in old formalin-fixed and paraffin-embedded (FFPE) gastrointestinal biopsies provides a means for such studies but is still challenging. Here we report a method based on laser capture micro-dissection and modified Roche 454 high-throughput pyrosequencing to obtain metagenomic profiles of Helicobacter pylori. We applied this method to two 15 year old FFPE biopsies from two patients. Frozen homogenized biopsies from the same gastroscopy sessions were also available for comparison after re-culture of H. pylori. For both patients, H. pylori DNA dissected from FFPE sections had ~96.4% identity with culture DNA from the same patients, while only ~92.5% identity with GenBank reference genomes, and with culture DNA from the other patient. About 82% and 60% of the predicted genes in the two genomes were captured by at least a single sequencing read. Along with sequences displaying high similarity to known H. pylori genes, novel and highly variant H. pylori sequences were identified in the FFPE sections by our physical enrichment approach, which would likely not have been detected by a sequence capture approach. The study demonstrates the feasibility of longitudinal metagenomic studies of H. pylori using decade-preserved FFPE biopsies.

Show MeSH
Related in: MedlinePlus