Bowtie Output "Not Enough Fields"

Question

[GPR]

Hello, I have ran Bowtie with the following command-line: 'bowtie -p 16 -q -n 3 -k 1 -m 1 --best --strata -S BowtieIndex *.fastq >& output.sam &' When running CleanSam.jar to remove overhanging reads, I get the following error: "Exception in thread "main" net.sf.samtools.SAMFormatException: Error parsing text SAM file. Not enough fields" I do not get this error when I align my reads with BWA or TopHat Any suggestions on how to fix this one? Thanks, G.

[Steve Lianoglou]

Have you tried skimming through some of your output.sam file to see if something looks blatantly wrong?

[GPR]

I did actually, nothing that caught my eye. This is a snapshot. << HWI-ST974:67:C0545ACXX:2:1101:14321:25195 0 chr2 242170185 255 76M * 0 0 CCATGATTTTGCGAATGGCTTTGCCGCGGGCACCAATGATGCGGGCGTGAACGCGGTGGTCCAGCGGGACGTCCTC CCCFFFFFHHHHHIJJIGIIJFIHIIIIGIIJIEHIFHHHHHFFDBB;<B?C?BDD8B@-:@CDCB5;@;9;5?@@ XA:i:0="" MD:Z:76="" NM:i:0<="" p="">

>

[GPR]

When I try to convert the same SAM file to BAM with "samtools view -bT genome.fa input.sam > output.bam" I get the following error message: " reference 'HWI-ST974:67:C0545ACXX:2:1101:11263:2144' is recognized as '*'. Parse error at line 4698: unmatched CIGAR operation"

[Philipp Bayer]

The "something is recognized as *"-error in samtools happens when you use a genome-reference (in your case, genome.fa) that doesn't include all the references that the sam-file input.sam lists.

It looks like you mistakenly aligned the reads to another set of reads, as the reference-name "HWI-ST974:67:C0545ACXX:2:1101:11263:2144" is a standard name for a read and not a chromosome or anything else.

Looking at your original command, you missed inserting the reference. Here's the fixed command:

bowtie -p 16 -q -n 3 -k 1 -m 1 --best --strata genome.fa -1 first_reads.fastq -2 second_reads.fastq > output.sam &

If I'm not mistaken, --best, --strata and --k 1 don't really make that much sense together. --k 1 just reports one alignment, while --strata tries to report all alignments that fall into the best stratum. What are you trying to achieve?

[matted]

Incidentally, this usage does make sense. The key is the -m 1 (which supercedes -k 1), saying that reads with more than 1 match are not reported. best and strata control the matching criteria: strata says that to count as a match, the alternate match must be in the same alignment stratum.

The relevant snippet from the bowtie documentation is: 'Intuitively, the -m option, when combined with the --best and --strata options, guarantees a principled, though weaker form of "uniqueness." A stronger form of uniqueness is enforced when -m is specified but --best and --strata are not.'