Another approach for comparison of metagenome datasets is fragment recruitment of metagenomic sequences related to selected genomes of reference species.
Molecules in the current dataset were fragmented into six fragments (R1, R2, R3, R4, R5, R6 .Group based QSAR Models were derived using Multiple linear regression (MLR) analysis and selected on the basis of various statistical parameters.
In addition, under the assumption that the chimpanzee and human reference sequences are representative of a randomly sampled individual in the genotype dataset, ePCR fragment lengths would be expected to lie within the union of the chimpanzee and human PCR fragment length ranges.
The molecules in the current dataset were fragmented into seven functional groups fragments (R1, R2, R3, R4, R5, R6 and R7).
Surprisingly, 51 of the recA homologs and 73 of the rpoB homologs identified by HMMER3 in the 1000 bp dataset were fragments sampled from outside the locus of these genes.
The size of the dataset (2312 fragments) is sufficiently large for statistical analysis.
As dataset for fragments of vectors and adaptors the Univec and Univec_core from NCBI were used.
To create a query dataset, 1000 fragments of lengths 200 and 1000 bp were sampled from random locations in each of the 334 microbial genomes.
In addition, we also generated a tenfold random resampled dataset of fragments of human chromosome 22 with distribution of lengths identical to those of the previously described observed dataset (see Materials and Methods, Fig. S1(see Additional file 1)).
These datasets have mean fragment lengths (114 nt and 140 nt, respectively) that are shorter than twice the read length (75 nt), meaning that most pairs of reads in these datasets will overlap at their 3-prime ends.
