Table 3. Summary of the computational processing of 454 pyrosequence data.

Step	Process	Program	Program source (URL or e-mail)	Calculator machine	# of query seqs.	Output	Time
1	Barcode sorting	split_libraries.py	QIIME (http://qiime.org/scripts/split_libraries.html)	Windows 64 bit PC*	1,583,218**	106 to 8618 reads per subject	10 min
2	OTU clustering	pick_otus_through_otu_table.py	QIIME (http://qiime.org/scripts/pick_otus_through_otu_table.html)	Windows 64 bit PC	131,768	15,578 redundant seqs.	17 min
3	de novo chimera check	de novo Uchime	Mothur 1.25.1 (http://www.mothur.org/wiki/Download_mothur)	Windows 64 bit PC	15,578	7,200 chimeric seqs.	6 min
4	DB chimera check	DB Uchime	Mothur 1.25.1 (http://www.mothur.org/wiki/Download_mothur)	Windows 64 bit PC	15,578	7,852 chimeric seqs. (6,612 chimeric seqs.)***	7 hrs
5	Sequence search up to genus level	RDP Classifier	RDP II (http://rdp.cme.msu.edu/)	RDP host computer	8,966	9 phyla–109 genera	10 min
6	Sequence search in species level	RDP Seqmatch	RDP II (http://rdp.cme.msu.edu/)	RDP host computer	8,966	3,992 seqs identified with know species	3 hrs
7	Data processing of RDP Seqmatch	SeqmatchQ400	Kyushu Univ(nakayama @ agr.kyushu-u.ac.jp)	Windows 64 bit PC	8,945	276 species	30 min

*Intel Core i7-3930 K CPU (3.20 GHz). **Batch sequence data included all sequences from 2 x half PicoTiterPlate regions, which contained 256 samples including non-Singaporean samples. ***The number of chimeric sequences determined by taking into account both de novo and DB chimera checks.