A more stringent clustering analysis of the genes of clusters 11 and 12 resulted in four and three subgroups respectively (Fig. 3).
Alternative TFBSs and palindromes were found for ESrrb (Estrogen-related receptor beta) when a larger word width (K = 11) was employed to capture more sophisticated co-occurrence and a more stringent clustering threshold was set to discriminate sub-patterns.
The user may choose a more stringent clustering criterion, i.e., at least two EST clones must be shared for linking two contigs together in order to reduce potential false linkages (similar to other practice, e.g., [ 12]).
The Inflation value is an indirect indication of "tightness" of clustering, such that the higher the inflation, the more stringent the clustering [ 24, 25].
Downstream 10K clusters are enriched in unannotated 3′ UTRs (Licatalosi et al., 2008 ); see also Table 2. (F ) Distribution of BC = 4 clusters by peak height for both nuclear and cytoplasmic HITS-CLIP; more stringent cytoplasmic clusters show enrichment in 3′ UTR.
Using this more stringent gene set, clustering further improved the separation of samples in clinically and pathologically defined groups, especially for the AD, PSP and control samples (see Figure 2A).
In an attempt to prevent the formation of chimeras derived from the assembly of sequences belonging to close paralogous genes, the assembly project reported here involved slightly more stringent conditions during clustering.
Many of the families, including the two largest, have little significant sequence identity, for any two proteins, in pairwise comparisons; however, this result did not change when we used more stringent parameters for clustering.
This voxel-level cluster-determining threshold has been used in the previous studies (Takeuchi et al. 2010b, 2011e, 2013a), and the validation study (Hayasaka and Nichols 2003) showed this threshold does not cause anti-conservativeness and if anything it seems, it leads to slight conservative results compared with more stringent voxel-level cluster-determining thresholds.
To minimize the potential errors in the detection of SNPs caused by sequencing errors, putative high-quality SNPs were further identified by using more stringent criteria (only clusters with at least four EST sequences and an SNP mutated within at least two ESTs were selected) and no indel data were included in the analysis (Supplementary Table S10).
