make.contigs issues with read numbers

[khoughton]

@khoughton

I am running Mothur v.1.37.6 with a Windows 7 PC.

I have just received paired end 16S mi-seq reads from Argonne and I am trying to make.contigs and demiltiplex the data. I received three files, the forward reads in one R1 file, reverse reads in one R2 file, and the index file (I1), with the sequences still contain the barcode but not the primers.

The first step I took was to create the reverse compliment for my oligos file. I followed Pat’s instructions using the reverseComp function in R. I then tried to run this oligos file with my data. My oligos file had “barcode” in the first column, the forward reading barcode in the second column, the reverse compliment barcode generated from the reverseComp in the third column, and the sample ID in the fourth column. First two lines are below:
barcode TAGGACGGGAGT TAGGACGGGAGT GoM_8045
barcode AAGTCTTATCTC AAGTCTTATCTC GoM_8046

Then I ran the following command:
mothur >
make.contigs(ffastq=Run1_16S_R1_001.fastq, rfastq=Run1_16S_R2_001.fastq, findex=Run1_16S_I1_001.fastq, oligos=Run1_16S_reverse_oligos.txt)

Using 1 processors.
Making contigs…
Done.
It took 22175 secs to process 16688339 sequences.

However, my groups file was empty and when I ran summary.seqs, it said my number of sequences had almost doubled, 32,930,404.
mothur >
summary.seqs()
Using Run1_16S_R1_001.trim.contigs.fasta as input file for the fasta parameter.

Using 1 processors.

Start End NBases Ambigs Polymer NumSeqs
Minimum: 1 243 243 0 3 1
2.5%-tile: 1 252 252 0 4 823261
25%-tile: 1 253 253 0 4 8232602
Median: 1 253 253 1 4 16465203
75%-tile: 1 253 253 1 5 24697804
97.5%-tile: 1 377 377 15 6 32107144
Maximum: 1 502 502 115 251 32930404
Mean: 1 260.384 260.384 1.95789 4.62585

of Seqs: 32930404

I feel my mistake is in my oligos file format and my make.contigs command. Is there a solution to this?

@pschloss

Can you post this as a new thread? FWIW, I think you might want to use NONE in place of the forward barcodes.

barcode NONE TAGGACGGGAGT GoM_8045
barcode NONE AAGTCTTATCTC GoM_8046

I posted this in Demultiplexing MiSeq Paired Reads in Oct '16. I am reposting it here, per the Pat’s request.

I Included NONE in the third column, since the samples were forward barcoded and it ran! I have been processing multiple sequencing runs and run into an issue. From my understanding, when mothur states “it took XXX secs to process XYZ sequences” that includes all the forward and reverse sequences (from the R1 and R2 files). Once the make.contigs is completed and we view the summary.seqs() we should only have half the number of sequences. If I am looking at this wrong please correct me, but I can’t figure out why some of my sequencing runs create more than half the number of total sequences after making the contigs. Examples below:

Run1 16S
It took 18107 secs to process 16,688,339 sequences.
Total of all groups is 6,738,318
(40%)

Run2 16S
It took 17512 secs to process 16,557,004 sequences.
Total of all groups is 12,642,768
(76%)

Run3 16S
It took 42626 secs to process 15,482,172 sequences.
Total of all groups is 5,720,164
(40%)

Run1 18S
It took 7417 secs to process 13,045,219 sequences.
Total of all groups is 10,458,033
(80%)

Run 2 18S
It took 10243 secs to process 16,636,473 sequences.
Total of all groups is 7,408,169
(45%)

Run 3 18S
It took 9785 secs to process 14,528,014 sequences.
Total of all groups is 12,350,637
(85%)

[pschloss]

From my understanding, when mothur states “it took XXX secs to process XYZ sequences” that includes all the forward and reverse sequences (from the R1 and R2 files).

That’s “process XYZ [pairs of] sequences”. So you wouldn’t expect to recover <50%. Does this help?