Article Figures & Data
Figures
- FIG. 1.
Sequence conservation as a function of alignment position for the 489,840 sequences in version 9.59 of the RDP. The x axis shows the position in the alignment as numbered by the E. coli 16S rRNA gene. The y axis shows the Shannon sequence entropy (see Materials and Methods), a widely used measure of conservation in multiple sequence alignments (23). Highly conserved positions within the alignment have a sequence entropy close to zero and hence are shown toward the top of the y axis. Positions of the hypervariable regions V1-V3 and V6-V7 are derived from Chakravorty et al. (3).
- FIG. 2.
Number of sequences assigned at the family classification level by the RDP classification algorithm to different sequences for the V1-V2, V6, and V6-V7 primers plotted against the number of 16S sequences assigned to the whole-genome shotgun sequence set. One has been added to each count to allow the data to be shown on a log-log plot.
- FIG. 3.
Across taxonomic levels, the results of a linear regression on log-transformed data between sequences generated by PCR targeting the 16S rRNA gene and 16S sequences culled from our whole-genome shotgun sequence set. Assignments are by the RDP classification algorithm as in Fig. 2. For each sequence set, two separate regressions were constructed, one for the rare biosphere (circles) with taxa seen 10 or fewer times in that sequence set's PCRs and one for a common biosphere (squares) with taxa seen 11 or more times in the PCR reactions (Fig. 1, gray lines). The top panels show the −log10 of the P value of the null hypothesis that the slope of the regression equals zero. The middle panels show Pearson's R values while the bottom panel shows the number of taxa for which classifications are made. Note that a significant P value (top panel) can be produced by either a negative or positive correlation.
- FIG. 4.
Comparison of classifications at the family level made by JGast and the RDP classification algorithm.
- FIG. 5.
Regressions across classification levels on log-transformed data showing the comparison between the RDP classification algorithm and JGast. Two regressions were constructed for each comparison: one for a common biosphere in which a taxon was observed 11 or more times under either JGast or RDP (squares) and one for a rare taxon in which fewer than 11 taxa were observed under both classification schemes.
Tables
- TABLE 1.
Primers used in initial 16S rRNA gene experimentsa
Primer Sequence Source or reference 27F gcctccctcgcgccatcagMGAGTTTGATCCTGGCTCAG 15 337R gccttgccagcccgctcagGCTGCCTCCCGTAGGAGT Fig. 1 967F gcctccctcgcgccatcagCAACGCGAAGAACCTTACC 25 1046R gccttgccagcccgctcagCGACAGCCATGCANCACCT 967F gcctccctcgcgccatcagCAACGCGAAGAACCTTACC 25 1177R gccttgccagcccgctcagCGTCATCCCCACCTTCCT Fig. 1 ↵a The lowercase sequence represents the 454A and 454B primers while the uppercase sequence represents conserved regions of the 16S rRNA gene.
- TABLE 2.
Sequences obtained via pyrosequencing of our 20 March 2007 aeration basin wastewater sample
Primer position No. of sequences generated Read length (bp)a Fraction of sequences with correct 5′ primer sequence Fraction of sequences with no N residues No. of sequences remaining after filteringb 27F-337R 15,141 228.2 ± 57.0 0.96 0.88 11,316 967F-1046R 15,124 92.7 ± 13.6 0.95 0.95 12,277 967F-1177R 18,873 199.9 ± 41.6 0.96 0.76 11,645 Whole genome 378,601 250.4 ± 29.11 ↵a Read lengths are given as means ± standard deviations.
↵b The filters removed any sequences where the reads (including nucleotides derived from the primer sequence) were below a size cutoff (70 bp for 967F-1046R and 150 bp for all others), did not start with the exact 5′ primer sequence, or had any N residues anywhere in the sequence.
- TABLE 3.
Number of sequences classified by RDP analysis scheme with a confidence score of >80%a
Region No. of sequences assigned to: Root Kingdom Phylum Class Order Family Genus Species V1-V2 11,316 11,313 10,646 9,824 6,979 4,647 2,287 205 V6 10,950 10,778 7,386 6,410 4,862 4,257 1,780 702 V6-V7 11,645 11,643 10,144 8,954 6,544 5,563 2,554 871 Whole genome 647 435 148 136 111 101 62 5 ↵a See Fig. 2 and 3.
- TABLE 4.
Number of sequences classified by JGasta
Region No. of sequences assigned to: Root Kingdom Phylum Class Order Family Genus Species V1-V2 11,316 11,316 11,137 10,968 10,104 9,513 6,907 424 V6 12,277 11,804 11,451 11,375 9,938 9,570 6,335 524 V6-V7 11,645 11,644 11,367 11,291 9,962 9,753 6,968 1,277 ↵a Assignment of sequences required that there be at least 85% sequence identity between the query sequence and the reference sequence in the Greengenes file and that the nearest neighbor(s) be classified with at least an 80% RDP confidence at the given taxon.
Supplemental material
Files in this Data Supplement:
- Supplemental file 1
-
Unfiltered Fasta sequences produced by our 454 run for all primer sets (File S1).
Zipped FAS files, 1.8 MB. - Supplemental file 2
-
Java code for JGast. (File S2).
Zipped JAVA files, 274K. - Supplemental file 3
-
Assignments generated by the RDP algorithm at a confidence score of 80% (File S3).
Zipped MS Excel files, 5K.
- Supplemental file 1
-
Unfiltered Fasta sequences produced by our 454 run for all primer sets (File S1).
