Sentence examples for molecular apocrine samples from inspiring English sources

We apply SAM to the index cohorts and identify a 346-probeset signature at 0% false discovery rate to predict molecular apocrine samples (see Additional file 3).

Significance Analysis of Microarrays (SAM) was performed on the normalized Doane et al. [ 7] and Farmer et al. [ 8] data individually to identify top 100 probesets that classify between the molecular apocrine samples and the remaining samples.

We note that median-centering per probeset by institution also results in statistically significant separation (p < 0.0001, see Additional File 1 -Figure S1) with 13 of 16 (81%) molecular apocrine samples clustering together in the HC dendrogram (see Additional File 1 -Figure S2) compared to 15 of 16 (94%) samples using XPN.

SAM was also performed on the combined Doane et al. and Farmer et al. subset of the cross-study normalized, five-cohort data to identify a gene signature with 0% false discovery rate for classifying molecular apocrine samples from the remaining samples, and identifying similar molecular trends in the remaining data.

These results show a natural demarcation in the larger ER- dataset where the 22 molecular apocrine sampes in the index cohort along with an additional 46 samples in the expanded cohort separate from the rest of the data (see Figure 7).

We evaluate our second proposed criterion for determining molecular equivalence by using Significance Analysis of Microarrays (SAM) [ 28] to identify the top 100 statistically significant probesets in each of the index cohorts (after normalization) that differentiate the hypothesized molecular apocrine phenotype from the remaining samples.

In order to initialize the BCRI strategy, we use a classifier method called See5 (Rulequest, St. Ives, Australia) to build a prediction model from the normalized gene expression data for classifying the molecular apocrine phenotype from the remaining samples in the index cohorts.

The non-BLBC tumors in our TNBC dataset mainly represent samples of the "molecular apocrine" type (16.5%), which demonstrates the inverse bimodal distribution as the basal-like metagene, and a relatively small group of "claudin-low" tumors (6.3%).

Therefore, we expand our molecular apocrine gene expression data with ER- samples from Ivshina et al., Rouzier et al., and Sotiriou et al., bringing our total to 199 [ 7, 8, 16- 18].

The dendrogram from the HC results shows 12 of 16 (75%) samples defined previously as molecular apocrine in a single cluster (p < 0.0001).

The cohort from Farmer et al. [ 8] includes 22 ER- breast carcinoma samples with six classified as molecular apocrine.