Hello everybody,
I am sorry if this topic is repeated, but I haven’t been able to find some help in other threads.
I am running shhh.flows now for one of the four regions I divided the plate, and I am not sure if it running OK. We are using Putty connecting to a Linux (16 processors) hosted on my local computer Win-7 without using MPI. It has been running for 29 hours now. At the beginning it ran quick, but It is in the same sample since yesterday, taking 150s per cycle…is that normal?
All comments, suggestions are wellcome
A couple things…
You have four regions, but how many barcodes? - are you running trim.flows first to separate the flow grams by sample?
Are you using the default flowgram parameters? If not what are you using?
Pat
Hi Pat,
and thanks for your reply.
We are using 100 different barcodes per run. I use the same barcodes for all the regions, except 4-5 different barcodes in each region in order to distinguish them and indeed I ran trim.flows before shhh.flows to separate the flowgrams by sample.
It took me 2 days and a half to run only one region of the plate. The other ones ran in a few hours. I am running again a duplicate of this region in a different run, and it happens the same. As I told you, this command is running fast at the beginning, and then gets held up in one of the samples. I guess this is because there is a problem with that sample?
Iam using the next parameter:
mothur> shhh.flows(file=er5.flow.files, processors=16)
Thank you in advanced
So each sample/barcode will run separately through shhh.flows. If you have 99 small samples and 1 big sample, the big sample will suck up a lot of time.
What are the parameters you are using in trim.flows? Can you post the entire command?
These are the parameters I am using in trim.flows:
mothur> trim.flows(flow=er9.flow, oligos=er9.oligos, pdiffs=2, bdiffs=1, minflows=360, maxflows=720, processors=2)
I am trying to use mpi to run shhh.flows with datas from the third run.
The minflows and maxflows in trim.flows must equal the same thing. If you look at http://www.ncbi.nlm.nih.gov/pubmed/22194782 you’ll see that we showed that if you use the 360/720 strategy you get very little denosinging, but that if you use 450/450 you get very good denoising. Furthermore, it will run much faster.
Pat