User case studies

Here we describe some interesting user case studies, that might help future users if they have a dataset similar to the ones tested here and wonder how DBGWAS would behave in their datasets.

We thank the users a lot and we appreciate their time for providing us the description and the output of DBGWAS in these examples. All credits go to the users.

If you want your case study to be featured here, please send the output and some description of your dataset by opening an issue at https://gitlab.com/leoisl/dbgwas/issues or sending it by mail to leandro<dot>ishi<dot>lima<at>gmail<dot>com .

Mass pen

• Author and credits: John Lees
• Description: This example investigates how DBGWAS copes with complex regions in genomes. More specifically, in this example, the author investigated how DBGWAS copes with close SNPs, which may split up the unitig graph into low frequency paths and lose association power (a known problem with k-mers). DBGWAS was run for beta-lactam resistance in S. pneumoniae, which is caused by mosaic alleles of three pbp genes. DBGWAS, with default settings, returned five hits, which using blastn of the sequences of the most significant nodes of the graph against a reference genome could be determined to be pbp2b, pbp2x, pbp2x, pbp1a, and pbp2x (ordered by q-value). This shows that DBGWAS can still find the relevant determinants even in highly variable regions.
• Dataset details: http://pyseer.readthedocs.io/en/master/tutorial.html
• Resources usage: 7.2 Gb RAM and 2.6 hrs CPU using 4 threads.
• Output availability: here