ng ERa or PR and all cases showing an HER2 score of less than 2 were considered negative. For CK5/14 and EGFR a semiquantitative score as well as the relative percentage of positive tumor cells was calculated. Informed written consent was obtained from each patient, and the study was approved by the Institutional Review Board at Hannover Medical School. For each patient, Rapastinel manufacturer genomic DNA was isolated from peripheral white blood cells using standard phenol-chloroform extraction. Sequencing data were analysed with NextGENe 2nd Generation Sequencing Software v.2.2.1. In brief, raw data were converted to FASTA files and were aligned to BRCA1 and BRCA2 gbk files from the human reference sequences. Only reads over 25 bases were converted, and reads were rejected if they contained more than 3 uncalled bases. Alignment was performed with a required matching of over 85% within more than 50 bases. This yielded an average of 2.35 million converted reads per sample, and about 95% of the reads could be matched. The average read length per sample was 487 bases, and the average coverage per sample was 74-fold. The average coverage per exon was above 30fold except for three amplicons that were covered less than 20-fold; these three exons and the missing exon 22 of BRCA1 were manually resequenced using BigDye Terminator Cycle sequencing with exon-flanking intronic primer pairs. For the others, mutation filters were set to exclude mutations with a percentage less than 10% or less than 3 counts, and to exclude homopolymer indels with a forward/reverse balance less than 0.1. Two regions were further inspected manually in each sample as they were largely represented by only one sequenced strand. All identified mutations, apart for common polymorphisms or known synonymous variants, were finally validated by conventional Sanger sequencing using BigDye chemistry and a 3100 Avant Genetic Analyser. PALB2 Analysis All exons of PALB2 were scanned for mutations by highresolution melting analysis as previously described. In brief, PCR amplifications were set up in the presence of the EvaGreen dye, and high-resolution melting analysis was performed on the Rotor-Gene 6000 realtime PCR machine. Melting profiles were evaluated using the Melt Curve Analysis tool of the Rotor-Gene 6000 Series Software Version 1.7. All samples with suspicious melting behaviour were then subjected to direct sequencing to identify the underlying substitution using BigDye chemistry and a 3100 Avant Genetic Analyser. BRCA1 and BRCA2 Analysis Target-specific primers were designed by Fluidigm Corp using Fluidigm primer service program with the following recommendations: Tm range of 5961uC, max of homopolymer is 3 and GC% less than 65%. Common sequence tags were added to forward and reverse primers for Access Array amplicon tagging experiments. 77 primer pairs were designed and validated to cover all exons except exon 22 in BRCA1. The exons of BRCA1 and BRCA2 were then amplified from triple-negative breast cancer patients to receive 40 pools of 77 amplicons. For this purpose, each genomic DNA sample was normalised to a concentration of,50 ng/ml and loaded onto an Access Array, a microfluidic array in which a PCR was performed with nested primer pairs. Each primary primer pair contained the templatespecific sequence and a tag sequence. Each secondary primer pair with sample contained the anti-tag sequence, a sample-specific unique barcode, and the 454 adaptor sequence. PCR products were harvested from