illumina pyrosequencing

2021/05/21

For example, the high coverage of indigenous communities provided by NGS has made it possible to quantitatively assess the impact of diet on human gut microbiota [8] and the diversity of metabolic pathways within marine planktonic communities [9]. 2A, inset). Citation: Luo C, Tsementzi D, Kyrpides N, Read T, Konstantinidis KT (2012) Direct Comparisons of Illumina vs. Roche 454 Sequencing Technologies on the Same Microbial Community DNA Sample. Analyzing raw (not assembled) reads, as opposed to assembled contigs, is typically restricted to cases where community complexity is too high or to specialized studies that aim to determine in situ abundance and/or population genetic structure and recombination [4], [10]. For instance, we noted that homopolymer-associated, single-base errors affected 1% of the protein sequences recovered in Illumina contigs of 10 coverage and 50% G+C; this frequency increased to 3% when non-homopolymer errors were also considered. The Emory Genome Center acknowledges the Georgia Research Alliance and the Atlanta Clinical and Translational Sciences Institute for funding for major equipment purchases. 1B. The results from metagenomic samples were further validated against DNA samples of eighteen isolate genomes, which showed a range of genome sizes and G+C% content. The frequency of single-base errors decreased with higher coverage of the corresponding contigs, i.e., the frequency dropped by about ten fold in contigs with 20 coverage relative to contigs with 2 coverage, reaching a plateau at about 20 coverage. Our work also provides a methodology for evaluating and comparing metagenomic data from NGS platforms. (A) A's and T's contribute significantly more homopolymer errors than C's and G's. Graph shows the variation observed in assemblies from different (replicate) datasets of the same genome; red bars represent the median, the upper and lower box boundaries represent the upper and lower quartiles, and the upper and lower whiskers represent the largest and smallest observations. 4), despite the fact that reads were trimmed based on the same quality standard prior to the analysis. We found that about 90% of the Roche 454 unique contig sequences overlapped with Illumina contig sequences (Fig. 454 roche sequencing platform generation biogene Homopolymer disagreements between the sequences in the alignment were identified and counted using a custom Perl script (the same approach was applied to the isolate genome data as well). We aligned the assembled contigs from 9 Illumina and 8 Roche 454 assemblies from JGI data for the same genome against the TIGR reference assembly and calculated base call error rate and gap open error rate as described above for JGI genomes. We found a strong linear correlation (r2>0.99) between the Roche 454 and Illumina data with this respect (Fig. Although our metagenomic analysis is based on a single community sample, we believe it is robust and informative. https://doi.org/10.1371/journal.pone.0030087.g004. Shared reads were defined as those that mapped on reads of the other dataset using Bowtie with default settings [25]. KyrpidesN, No, Is the Subject Area "DNA sequencing" applicable to this article?

It is possible that the remaining 10% of the contig sequences might have been different because of imperfect or uneven splitting of the original DNA sample into the two aliquots sequenced and the fact that the diversity in the sample was not saturated by sequencing (estimates based on rarefaction curves using raw reads indicated that we sampled about 8085% of the total diversity in the Illumina data). Graphs show the calculated base call error rate (A) and gap open error rate (B) for each comparison (figure key). (2012) Some of our results (e.g., assembly N50 comparisons, Fig. ReadT, These results were attributable to a higher number of (artificial) frameshifts, caused by homopolymer-associated base call errors, present in the Lanier.454 versus the Lanier.Illumina assembled sequences. Due to frameshifts caused primarily by homopolymer-associated errors in the derived consensus sequence of the contigs, genes from Roche 454 assembly had fewer complete matches in the NR database relatively to their Illumina counterparts (inset; results are based on a total of 72,709 gene sequences annotated on contigs that were shared between the two assemblies and were longer than 500 bp). No, Is the Subject Area "Metagenomics" applicable to this article? The 95% identity cut-off was used to accommodate the maximum sequencing error observed in raw reads of an isolate genome (about 5%); other cut-offs are not as appropriate as the one used above and were not evaluated. Nevertheless, about 1% of the total genes recovered in the Illumina assembly contained homopolymer-associated sequencing errors and this number increased to about 3% when non-homopolymer-associated errors were also taken into account (for contigs showing 10 coverage, on average). Collectively, our results should serve as a useful practical guide for choosing proper sampling strategies and data possessing protocols for future metagenomic studies. 3), low G+C% genomes sequenced with this platform may have 20% or more genes with frameshift errors whereas the Illumina platform is not affected as much by the G+C% of the sequenced DNA (Fig. https://doi.org/10.1371/journal.pone.0030087.g003. Velvet was used to assemble each of these Illumina datasets with K-mer set at 31. In order to account for possible biases introduced by uneven genus abundance and provide statistically robust estimates, we employed a Jackknifing resampling method. 4).

LuoC, The alignments were used to count frameshift errors separately for each Illumina or Roche 454 dataset. Yes The average G+C% content of the metagenome was 47.4%; thus, our results are not simply attributable to higher abundance of A's and T's in the metagenome.

Note that Illumina assemblies recovered a significantly larger fraction of the reference genome than Roche 454 assemblies (two tailed Whitney-Mann U test p-value=0.014), which is consistent with the results from the metagenomes (Fig. Yes Given that the single-base error of individual reads was comparable between Lanier.454 and Lanier.Illumina (0.5% per base), our results reveal that the lower single-base error rate of Lanier.Illumina contigs (3% vs. 4.5% for Roche 454, counting homopolymer- and non-homopolymer-associated errors) is primarily due to the higher coverage obtained. (B) Error rate (as a percentage of the total genes evaluated, y-axis) increases as homopolymer length increases (x-axis). Contigs were defined as shared between the assemblies of the Lanier.454 and Lanier.Illumina data when they shared at least 95% nucleotide sequence identity and overlapped by at least 80% of their length (for the shorter contig). These findings call for special attention in cases where the sequenced DNA (e.g., community or isolate genome) is of low G+C%. (B) Protein sequences annotated on raw (not assembled) reads matched genes in the reference assembly more frequently for the Roche 454 than the Illumina data. 2B). Red bars represent the median, the upper and lower box boundaries represent the upper and lower quartiles, and the upper and lower whiskers represent the largest and smallest observations. For instance, protein sequences called on Lanier.454 reads had 10% more Blastp matches to reference genes from the Lanier.454 assembly than did protein sequences from Lanier.Illumina reads against the Lanier.Illumina reference assembly (Fig. (A) Venn diagram showing the extent of overlapping and platform-specific raw reads between the Lanier.454 and Lanier.Illumina datasets (without assembly). We evaluated the type and frequency of errors in assembled contigs from metagenomic data using both a comparative and a reference genome approach. Performed the experiments: CL DT. Base call errors and gap opening errors were identified as discrepancies between the read sequence and the reference assembly sequence using a custom Perl script. Finally, our evaluations showed that the choices of parameters and amount of input sequence of the assembly did not have any dramatic effect on the quality of the resulting contigs for both Illumina and Roche 454 assemblies (Fig. succinogenes S85. The genomes were: Candidatus Pelagibacter ubique HTCC1062 (-Proteobacteria), Opitutus terrae PB901 (Verrucomicrobia), Polaromonas sp. Is the Subject Area "Genomics" applicable to this article? The same cut-off was used to map raw reads on contigs. School of Biology and Center for Bioinformatics and Computational Genomics, Georgia Institute of Technology, Atlanta, Georgia, United States of America, Affiliation To compare the quality of Illumina vs. Roche 454 contigs assembled from isolate genome data the following approach was followed: Illumina data for each genome was randomly sampled to form several technical replicate datasets, each of which provided about 100 coverage of the reference assembly, on average. We applied widely used protocols to assemble both sets of reads (see Materials and Methods for details), which substantially collapsed the Lanier.Illumina dataset into 57 Mbp of total unique sequences and the Lanier.454 dataset into 46 Mbp (Fig. 2) should be independent of the NGS platform considered and broadly applicable to short-read sequencing. For instance, derived assemblies overlapped in 90% of their total sequences and in situ abundances of genes and genotypes (estimated based on sequence coverage) correlated highly between the two platforms (R2>0.9). 3). The higher sequence error rate observed for the TIGR reference genome might be due to the different strain of F. succinogenes sequenced or differences in the sequencing platforms or the assembly protocols used by JGI and TIGR. It is, however, currently economically unfavorable to obtain similar coverage with the Roche 454 sequencer to the Illumina data (see Discussion below). 1C); 57.7% and 49.5% of the total reads in the Lanier.Illumina and Lanier.454 datasets, respectively, were singletons (i.e., remained unassembled). Competing interests: The authors have declared that no competing interests exist. Assemblies of isolate genome sequences (closed or high-draft) were downloaded from the NCBI RefSeq database (called reference assemblies for convenience); raw Illumina and Roche 454 sequencing reads were available through the Joint Genome Institute (JGI, www.jgi.doe.gov). It should be noted, however, that most of the previous error estimates and sequencing biases have been determined based on relatively simple DNA samples (e.g., a single viral genome) and thus, their relevance for complex community DNA samples remains to be evaluated. We used the isolate genome data to evaluate the effect of the parameters of the assembly on the quality of the contigs as follows: a series of assemblies were obtained for genomes of low (Arcobacter nitrofigilis, 28%), medium (Fibrobacter succinogenes, 48%), and high (Cellulomonas flavigena, 74%) G+C% content. These percentages were similar to those reported above based on the comparative method (the 3.3% of homopolymers that disagreed between the two datasets includes both Roche 454- and Illumina-specific homopolymer errors). These errors were not observed in the Illumina data, presumably due to both the high sequence coverage that greatly facilitated the resolution of homopolymer ambiguities and the less pronounced sequencing biases of Illumina (Fig. Department of Energy (DOE) Joint Genome Institute, Walnut Creek, California, United States of America, Affiliation Hence, the majority of non-homopolymer-associated errors remain challenging to model and thus, to correct. Even though read lengths increase as the technologies advance, they are still far shorter than the desirable length (e.g., the average bacterial gene length is 950 bp) or the read length obtained from traditional Sanger sequencing (1000 bp). correction. PLOS ONE promises fair, rigorous peer review, Yes Conversely, protein sequences annotated on Illumina reads more frequently matched to the wrong protein sequence in the reference assembly (mismatched genes) or did not match any reference gene (unmatched genes). Our previous study [17] as well as those of others [20], [21] reported high reproducibility of Illumina-based and 454-based DNA sequencing within the same community sample. We identified 0.4 million homopolymers (three identical consecutive nucleotide bases or more), of which 14 thousand (3.3% of the total) disagreed on length between the two assemblies, resulting in alternative amino acid sequences for about 7% of the total 72,709 gene sequences evaluated. Thus, the results reported for Illumina based on the metagenome of Lake Lanier (47 G+C%) should be also applicable to metagenomes with different G+C% contents.

To eliminate the possibility that our results were biased by the selection of reference genomes, we used the reference assembly of Fibrobacter succinogenes subsp.

Single-base sequencing errors increased by an average of 2% when non-homopolymer-associated errors were also taken into account for both platforms. We found that homopolymer errors affected 2.132.78% and 0.321.02% of the total genes evaluated for the Lanier.454 and Lanier.Illumina data, respectively (dividing by the average gene length, 950 bp, provided the per base error rate; range was estimated from 100 replicates using Jackknife resampling), despite the fact that sequencing error in the raw reads of the two platforms was comparable (0.5% per base, in our hands). 3), which is in agreement with previous results [5], [11]. PLoS ONE 7(2): Abundance was determined based on the number and coverage of the contigs, as described elsewhere [17]. For comparing gene calling accuracy on unassembled reads, we employed FragGeneScan [27] to predict genes on Lanier.454 and Lanier.Illumina reads using the 454 1% error rate model and the Illumina 0.5% error model, respectively. First, we examined disagreements in gene sequences annotated on contigs larger than 500 bp and shared between the Lanier.454 and Lanier.Illumian assemblies. We sampled 50% of the total homopolymers at random and estimated homolopolymer rate in this subset. succinogenes S85 genome sequenced at JGI were compared against the reference assemblies from the JGI and TIGR genome projects of Fibrobacter succinogenes subsp. Next generation sequencing (NGS) technologies, such as the Roche 454, Illumina/Solexa, and, to a lesser extent, ABI SOLiD, have been cornerstones in this revolution [5], [6], [7]. Wrote the paper: CL KTK. Discover a faster, simpler path to publishing in a high-quality journal. Conceived and designed the experiments: CL NK KTK. Evaluation of base-call error, frameshift frequency, and contig length suggested that Illumina offered equivalent, if not better, assemblies than Roche 454. In the reference genome approach, genes annotated in the Lanier.454 and Lanier.Illumina contigs were compared against their orthologs in publicly available genomes, and homopolymer errors were identified assuming the publicly available sequences contained no errors. here. No, PLOS is a nonprofit 501(c)(3) corporation, #C2354500, based in San Francisco, California, US, Corrections, Expressions of Concern, and Retractions, https://doi.org/10.1371/journal.pone.0030087, https://doi.org/10.1371/annotation/64ba358f-a483-46c2-b224-eaa5b9a33939. Among these genes, Roche 454 data appeared to have the wrong (artificial) sequence more often than Illumina data. Lanier.Illumina contigs were generally longer than Lanier.Roche 454 contigs, i.e., the assembly N50 (the contig length for which 50% of the entire assembly is contained in contigs no shorter than this length) was 1.6 Kbp versus 1.2 Kbp, respectively. Roche 454 sequencing quality is evaluated in panels A through D, which show: (A) base call error rate of individual reads (x-axis) for each genome evaluated (y-axis); (B) base call error rate (y-axis) plotted against the G+C% of the genome; (C) gap opening error rate of individual reads (x-axis) for each genome evaluated (y-axis); (D) gap opening error rate (y-axis) plotted against the G+C% of the genome. Department of Human Genetics, Emory University, Atlanta, Georgia, United States of America, Affiliations School of Biology and Center for Bioinformatics and Computational Genomics, Georgia Institute of Technology, Atlanta, Georgia, United States of America, The resulting datasets were 502 Mbp (Lanier.454) and 2,460 Mbp (Lanier.Illumina) in size; all our bioinformatic analyses and comparisons were based on these trimmed datasets. As noted above, similar gap opening errors were observed for the metagenomic reads from the two platforms and single-base accuracy was comparable between the two platforms (99.34% vs. 99.46% for the Lanier.454 and Lanier.Illumina metagenomic reads, respectively). Illumina GA II sequencing quality is evaluated in panels E and F, which show: (E) base call error rate of individual reads plotted against the G+C% of the genome; and (F) gap opening error rate of individual reads plotted against the G+C% of the genome. Thus, Roche 454 is advantageous with respect to gene calling when working with unassembled reads. 7). 2B). Illumina-specific unique contig sequences (16 Mbp) were more than three times as many as the Roche 454-specific ones (5 Mbp), and these additional contigs were attributed to the larger Illumina dataset rather than sequencing artifacts or errors. Therefore, the two platforms provided comparable in situ abundances for the same genes or genomes. This resulted in a set of 500 bp long sequence fragments, which were subsequently mapped onto the reference assembly using Blastn.

Protein-coding genes encoded in the assembled contigs were identified by the MetaGene pipeline [26]. We also found that the systematic single-base errors associated with GGC-motifs in Illumina data reported recently [16] represented only a minor fraction of the non-homopolymer-associated errors (0.015% of the total bases analyzed, consistent with the frequency reported in the original study). DT acknowledges the support of the Onassis Scholarship Foundation. The amount of Illumina and Roche 454 input sequence data was chosen so that the ratio of the two was similar to the ratio in the metagenomic analysis (2.5 Gb Illumina reads versus 500 Mbp Roche 454 reads, or 51).

Assembly parameters (primary and secondary x-axes) were evaluated for low (Arcobacter nitrofigilis, 28%; left), medium (Fibrobacter succinogenes, 48%; middle), and high (Cellulomonas flavigena, 74%; right) G+C% genomes. Yes In the former approach, we examined protein-coding sequences recovered in contigs longer than 500 bp that were shared between the Lanier.454 and Lanier.Illumina assemblies. Copyright: 2012 Luo et al. https://doi.org/10.1371/journal.pone.0030087.g006. Finally, in all genomes analyzed, Illumina assemblies consistently recovered a larger percentage of the reference genome than Roche 454 assemblies (two tailed Whitney-Mann U test p-value=0.014; Fig. The results reported represent averages from 100 iterations. broad scope, and wide readership a perfect fit for your research every time. Even when only a fraction of the total Illumina dataset was used in the analysis that was comparable to the size of the Roche 454 dataset (i.e., 500 Mbp), the derived Illumina assemblies were similar to those of Roche 454 (N50 values were 990 bp for Illumina and 1193 bp for Roche 454; Fig. NGS platforms produce millions of short sequence reads, which vary in length from tens of base pairs (bp) to 800 bp. We extracted the predicted gene sequences from the reads and the corresponding amino acid sequences were searched against the genes of the reference assembly of the same dataset using BLAT [28]. We compared the reads from the Lanier.Illumina dataset against the Lanier.454 dataset to identify the fraction of reads shared between the two datasets. View School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, Georgia, United States of America, Affiliation We assessed the advantages and limitations of the Roche 454 and Illumina platforms for metagenomic studies by sequencing the same community DNA sample with each platform. Second, we directly assessed homopolymer error rate against reference genomes from GenBank that represented close relatives (average amino acid identity >70%) of the microorganisms sampled in the Lanier metagenome. The results for the isolate genomes were based on Illumina input reads that were about 5 times as many as the Roche 454 input reads to provide a ratio that was similar to that of the metagenomic comparisons (51). This corroborated our estimated error rate in metagenomic data, i.e., that the Lanier.454 assembly had 7% more frameshift sequences than the Lanier.Illumina assembly (Fig.

The two platforms agreed on over 90% of the assembled contigs and 89% of the unassembled reads as well as on the estimated gene and genome abundance in the sample (Fig. Assemblies were obtained for each possible combination and the base call error and gap opening error of the resulting assemblies were determined as described for individual reads above. Noticeably, due to the inherent biases of the Roche 454 sequencing approach to produce more frameshifts in A and T rich DNA (Fig. Yes These patterns were not as pronounced in the Illumina data, indicating that Illumina errors were (more) randomly distributed than Roche 454 errors (see Fig. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Affiliation Reciprocal best matches (RBMs), when overlapping by at least 500 bp and showing higher than 95% nucleotide identity, were identified and re-aligned using ClustalW2 [31]. We would like to thank Chad Haase and Ryan Weil for their assistance with sequencing and Rachel Poretsky for critically reading the manuscript. Yes Individual reads were mapped against the assembled contigs using Bowtie [25] with default settings to calculate average contig coverage. Although the use of the TIGR reference assembly resulted in a slightly higher number of sequence errors for both Illumina and Roche 454 data, Illumina consistently showed a smaller number of sequencing errors and the relative error rate between the two platforms was similar to that based on the JGI genome data alone, independent of the reference genome used (Fig. 4, which is based on isolate genome data). Sequences shorter than 200 bp (Lanier.454) and 50 bp (Lanier.Illumina) after trimming were discarded. We also estimated the abundance of each contig shared between the two assemblies by counting the number of reads composing the contig, which can be taken as a proxy of the abundance of the corresponding DNA sequence in the sample [19]. 4, 5, 6 and Table 1). 2). KonstantinidisKT From the human gastrointestinal tract to the ocean abyss, whole-genome shotgun metagenomics is revolutionizing our understanding of the structure, diversity, and function of microbial communities [1], [2], [3], [4]. Six genomes that represented abundant genera in the lake metagenome were identified this way. (C) Assemblies were obtained from 502 Mbp of Roche 454 and 2,460 Mbp of Illumina data using established protocols. For convenience, we called the two sequence data sets Lanier.454 and Lanier.Illumina, respectively. Contributed reagents/materials/analysis tools: NK TR. We did not observed a significant difference in error frequency in contigs with higher than 20 coverage (standards on length and coverage for identifying error-prone Illumina contigs are defined in our previous study [18]). Although low coverage contigs (e.g., 1 to 5) are likely to contain a higher fraction of chimeric sequences than 0.2% according to our previous study [18], such contigs were rare in the results reported here, which included only contigs longer than 500 bp with average coverage 10 or higher (only about 3% of the contigs showed less than 5 coverage; Fig.