Mac/Windows Variation in the results of make.contigs

mothur bugs
1

Hi, I am new to mothur and would really appreciate some help.

I faced with one problem, which might be a bug. I tested exactly the same make.contigs commands, oligos files and Miseq fastq files on Mac and Windows versions of mothur 1.34.4 and faced the great variation in the results.

More particularly, this is my command:

make.contigs(ffastq=665fITS7_R2.fastq, rfastq=665fITS7_R1.fastq, oligos=ITS7ja4.oligos, pdiffs=2, processors=2)



next I’ve got the total of all groups as 3763 on Windows and 7423 on Mac, which looks suspicious. Is it a bug or I’ve missed something basic?
It is unlikely that the problem resulted from wrong formation of the oligos file, but just in case, I used NONE in the BARCODE line of the oligos file, since barcodes were trimmed previously, so the file looks like that:

primer GTGARTCATCGAATCTTT TCCTCCGCTTATTGATATGC fITS7R1_ITS4F1
primer WGTGARTCATCGAATCTTT TCCTCCGCTTATTGATATGC fITS7R2_ITS4F1
primer CCAGTGARTCATCGAATCTTT TCCTCCGCTTATTGATATGC fITS7R3_ITS4F1
primer GTGARTCATCGAATCTTT STCCTCCGCTTATTGATATGC fITS7R1_ITS4F2
primer WGTGARTCATCGAATCTTT STCCTCCGCTTATTGATATGC fITS7R2_ITS4F2
primer CCAGTGARTCATCGAATCTTT STCCTCCGCTTATTGATATGC fITS7R3_ITS4F2
primer GTGARTCATCGAATCTTT AGTRRTCCTCCGCTTATTGATATGC fITS7R1_ITS4F3
primer WGTGARTCATCGAATCTTT AGTRRTCCTCCGCTTATTGATATGC fITS7R2_ITS4F3
primer CCAGTGARTCATCGAATCTTT AGTRRTCCTCCGCTTATTGATATGC fITS7R3_ITS4F3
BARCODE NONE NONE q

2

This looks like a bug in the Windows version with processors > 1. Can you try running the Windows version with processors=1?

3

Thanks a lot, it works in the proper way with processors=1 on Windows

4

Hi

Does this mean that the whole protocol has to be carried out in windows with 1 processor, or just the make.contigs command.

I am having similar difficulties.

Thank you

5

That would be for just make.contigs, but this will be fixed in the next release.