recommended flow value for data from GS FLX+


I have sff files from GS FLX+.

I run “mothur >, flow=T)”
and next
I want to run “trim.flow(sample.flow, minflows= $$, maxflows=$$)”
but I don’t know the optimal flow value for GS FLX+.
when I run “trim.flow(sample.flow, minflows= 360, maxflows=720)” recommended values for GS FLX Titanium, the result file (.trim.flow) is empty.

I wonder the recommended flow values for GS FLX+


when I run “trim.flow(sample.flow, minflows= 360, maxflows=720)” recommended values for GS FLX Titanium, the result file (.trim.flow) is empty.

So first off… those are not the recommended values - 450 is the recommended value for both parameters for GS FLX Titanium.

Second, for GS FLX+, I’m going to assume that your sequencing provider used LA1, LA2, or LA3. If this is the case then you need to run…

trim.flow(sample.flow, minflows=1050, maxflows=1050, order=B)

This will produce a product that is about 570 bp and has an error rate on par with what we have seen in the past ~0.06%.

I really need to get off my duff and write this stuff up… In the mean time if you or anyone else would be interested in seeing a slide deck I prepared, I’d be happy to send it to you if you email me.

Thank you so much for your help :slight_smile:

I am analysing an FLX+ dataset at the moment. I found that I needed to use the B flow order but setting minflows and maxflows to 1050 yielded only c1000 sequences from a half-plate (there are 450K sequences in the raw data). The flow file header shows the number of flows as 1779. Do you have any further suggestions for settings?

How long is your fragment - I guess I should clarify my comment about 1050 - this is the largest value I would go up to. If your region is shorter than ~5¶0 bp, then you want to drop the number of flows. Also, if your sequence provider didn’t use LA1, 2, or 3, then all bets are off.


The analysis was LA3.

The amplicon was V1-V3 so normally just over 500 bp, depending on the species.

What figure do you suggest?

Many thanks for your help.

You might try 900

Thanks, I’m trying that.

shhh.flows is taking a while to run though (on a 8-processor Mac Pro). I suppose it’s the higher number of flows?

Yeah and I think there’s probably more sequences as well, but I forget how many more you get now.


I am analyzing 454 FLX+ data. The flow file indicates a total number of flows of 1779. The average length of the sequences is 700 bp.
How many min and maxflows should I try? I’ve seen recommendations from 900 to 1050 but I dont know how to decide and as I dont have a computer with very high capacity I’m not able to try several options and then decide.
Any recommendation would be appreciated!
Thank you all!

Do 1000.

Thank you!