Table 3. Sequence statistics by library.

	%GC	avg. read length	total reads	Mapped reads	Unmapped reads
					ambiguous hits to ref orgs	no hits to ref orgs
						hits NCBI	no hits to NCBI
Enz+Pyrosequencing	0.58	202	505962	495428	2354	8004	176
Enz+Sanger	0.56	578	11781	9054	364	272	2091
Additional Enz+Sanger	0.59	688	6542	5351	149	267	775
Glycerol Enz+Sanger	0.61	655	6186	4338	294	388	1166
EnzBB+Sanger	0.48	563	14418	11552	760	519	1587
Additional EnzBB+Sanger	0.56	665	2040	1258	39	266	477
Glycerol EnzBB+Sanger	0.52	699	1348	718	51	180	399
DNeasy+Sanger	0.6	568	14692	11865	759	592	1476
Additional DNeasy+Sanger	0.6	654	2625	1945	90	168	422
Glycerol DNeasy+Sanger	0.59	694	1726	1054	36	155	481

For each library, the average read length, percent G+C, total number of reads, and the numbers of mapped and unmapped reads are given. The unmapped reads fall into 2 categories: 1) those that have BLAST hits to our reference organisms, but cannot be mapped to a single organism because they have high sequence identity to more than one organism or because the sequence identity is below the 95% threshold; and 2) those that do not have BLAST hits to our reference organisms. Reads in the second category are further subdivided into reads that do hit other organisms in the NCBI non-redundant nucleotide database and reads that do not.