Negative start position bed file

Question

[chariko]

I aligned two samples (bwa; filter with samtools, removed duplicates with picard-tools, bamtobed and bed to bedgraph with genomeCoverageBed) to a genome file consisting in 4 genes and the first sample went fine but the second one when getting the bed file the start position for one of the genes was -1. I wonder that is not correct but I do not know why did I obtain such negative value.

Example:

"Wrong bed file for gene1"
gene1      -1      150      M00...      37         -
gene1      -1          150      M01...      37         +
gene2      1004      1115      M00...      60         -

I realized that -1 values in special cases, however, BEDPE (bed paired end) positions may be set to -1 to indicate that one or more ends of a BEDPE feature is unaligned.

http://bedtools.readthedocs.org/en/latest/content/overview.html

but I do not understand if that's my case and I really do not understand why did it align if it was unaligned....

Any clues?

[Devon Ryan]

Looks like an off by one error.

[chariko]

thanks for your answer, but... what do you mean off by one error?

would you consider to discard this result and that's all?

thanks again

[Devon Ryan]

Have a look at the wikipedia article, this is a pretty common type of program bug. You can probably just convert the -1 values to 0.

[Istvan Albert]

ok but if the zeros are off by one isn't it likely that the other rows are off by one as well ;-)?

[Devon Ryan]

Indeed :)

[Manvendra Singh]

Can't say, why its negative

Maybe you can just convert -1 to +1

Answer 1

[chariko]

After changing the data to 0 and 1, I decided to remove those reads because also the coverage was very low (3x) so I really do not belive the results.

Thanks all for your answers.