sff_extract: combining data from 454 Flx and Titanium data sets

[agroster]

Hi,
I am attempting to feed my 454 sequencing data sets into MIRA for assembly.
Here is my data, all in .sff files:
Flx unpaired reads (6 separate. sff files)
Titanium unpaired reads (3 separate .sff files)
Titanium paired-end reads (2 separate .sff files).

I can extract to fasta and xml with the individual sets (unpaired-only and paired-only), but the MIRA assembler requires all data to be combined into a single data file.

Does anyone know if this can be done with sff_extract? If so, can you explain to me how? The sff_extract webpage doesn't explain how to do this.

[Jose Blanca]

Have you tried to extract them toghether? Just like:

sff_extract 1.sff 2.sff

[agroster]

Thanks for your reply.

Since one set of my data is paired-end, i believe i need to force sff_extract to split up the reads by inducing the -l command to look for the linker. Also, mira requests an addition of the paired-end library size and stdev info, but I wouldn't want that applied to the unpaired reads.

I tried the cat command, but mira quit inexplicably when loading these files.

[maubp]

Originally Posted by agroster

I tried the cat command, but mira quit inexplicably when loading these files.

You can't just cat SFF files together - they have a complex header structure. It isn't really needed, but you can use the Roche tools (or 3rd party software) to combine SFF files, but they must all have the same number of flow cycles. In your case you have both FLX and Titanium runs so you can't do this.

In any case, don't you want to give MIRA two separate sets of 454 data
(the pairs and the unpaired reads)?

[agroster]

Originally Posted by maubp

You can't just cat SFF files together - they have a complex header structure. It isn't really needed, but you can use the Roche tools (or 3rd party software) to combine SFF files, but they must all have the same number of flow cycles. In your case you have both FLX and Titanium runs so you can't do this.

In any case, don't you want to give MIRA two separate sets of 454 data
(the pairs and the unpaired reads)?

I concatenated the separate .fasta, .qual, and .xml files that were generated post sff_extract for each type of data, not the .sff files before sff_extract.

I DO want to give MIRA two separate sets of 454 data, but I don't know how to do this (since MIRA looks for only one file name). Does anyone know how to do this?

[maubp]

Originally Posted by agroster

I concatenated the separate .fasta, .qual, and .xml files that were generated post sff_extract for each type of data, not the .sff files before sff_extract.

Concatenating FASTA and QUAL files is fine, but I don't think you should concatenate XML files together. MIRA may cope, but I would expect any XML validator to reject such a file.

Originally Posted by agroster

I DO want to give MIRA two separate sets of 454 data, but I don't know how to do this (since MIRA looks for only one file name). Does anyone know how to do this?

Have you read the "Walkthrough: combined unpaired and paired-end assembly of Brucella ceti" example in the MIRA manual?
http://chevreux.org/uploads/media/mi...tml#section_21

[agroster]

Ok, i now see the issue - I need to append my subsequent sff extractions using the "-a" option. i.e. do three subsequent extractions, each appending to the previous. I'll see if this works.

[BaCh]

Originally Posted by agroster

Ok, i now see the issue - I need to append my subsequent sff extractions using the "-a" option. i.e. do three subsequent extractions, each appending to the previous. I'll see if this works.

Ah, I'm too late. Actually, two extractions are enough. First all the unpaired, then all paired with -a.

Regards,
Bastien