For example, within our DNA variant annotation pipeline, key datasets include 1) Patient_Called_DNA_Variant_File, the file of sequenced DNA variants from a patient's tumor; 2) COSMICSubset, the GeneTrails-specific subset of COSMIC; 3) SNPSubset, the GeneTrails-specific subset of dbSNP; and 4) Final_Annotation_of_DNA_Variants, the final annotation file of the identified DNA variants.
Secondly, we tested the Parallel-META pipeline on Dataset 2. Finally, we analysed the comparison result and complete common phylogenetic tree by using Dataset 1.
Following initial processing using our bionformatics pipeline, the dataset had 1 261 575 non-host sequences for subsequent analyses.
This enables to launch new pipelines on datasets already stored in the system in order to answer new requests.
To assess BDS's scalability, we ran exactly the same pipeline on two datasets that vary in size by several orders of magnitude (Table 2): (i) a relatively small dataset (chromosome 20 subset, ∼2 GB) that would typically be used for development, testing and debugging and (ii) a high-depth whole-genome sequencing dataset (over 200× coverage, roughly 1.5 TB).
Moreover, by applying our pipeline to primate datasets of similar size we identify ≤1000 sites for each mismatch type.
Data processing, including Irreproducible Discovery Rate (IDR) analysis, was done using a uniform data processing pipeline for both datasets.
The tutorial "PCM with camb" (Additional file 2) reports the complete modelling pipeline for this dataset [28].
After running our pipeline on this dataset, the proportions of reads assigned to the respective reference genomes mirrored the sample compounds with high accuracy.
