These are chat archives for dereneaton/ipyrad

9th
Jan 2018
Isaac Overcast
@isaacovercast
Jan 09 2018 19:07
@jeremycandersen The problem is probably related to this: dereneaton/ipyrad#285
Isaac Overcast
@isaacovercast
Jan 09 2018 19:55
Read the comments in that ticket for more details, but the short version is there's a pe ddrad multiplex barcode protocol that sticks the second barcode in the sequence identifier line. The 3rad protocol actually puts the second barcode in the sequence, in a similar fashion to the first barcode. The only multiplex barcode protocol ipyrad supports is this exact 3rad protocol. Can you tell me the reference for the library prep you followed for this data? I made a ticket to support this but I can't make any promises about when it'll get done.
@nitishnarula You might try -t 4 to see how it goes. For ipyrad more cores is always better (in most cases), our parallelization strategy is very aggressive. As for whether 800GB is sufficient I would hope it is, but your dataset is very large, so it's hard to predict. You might try starting the run and then monitoring memory consumption to see if it gets pegged.