I am in the process of compairing different short read mappers for a task.
I want to use BBMap's "randomreads.sh" & "samtoroc.sh" to do that. (Version 36.84)
With BWA it worked like a charm. But I get strange results when I use bbmap.sh to map my simulated reads.
Commands are the following:
randomreads.sh ref=../pe_meta_contigs_minLength.fasta paired=f mininsert=200 maxinsert=1000 length=300 coverage=200 spaceslash=f
bbmap.sh ref=../pe_meta_contigs_minLength.fasta in=reads_B1_3938618x300bp_0S_0I_0D_0U_0N.fq.gz out=SE_bbmap_onSimulatedData.sam
samtoroc.sh in=SE_bbmap_onSimulatedData.sam reads=3938618 > roc.txt
Result:
ROC Curve for SE_bbmap_onSimulatedData.sam
minScore mapped retained truePositiveStrict falsePositiveStrict truePositiveLoose falsePositiveLoose falseNegative discarded ambiguous
46 92.9880 92.9880 0.0000 92.9880 0.0000 92.9880 7.0120 0.0000 0.0000
45 96.6744 96.6744 0.0000 96.6744 0.0000 96.6744 3.3256 0.0000 0.0000
44 98.1959 98.1959 0.0000 98.1959 0.0000 98.1959 1.8041 0.0000 0.0000
43 99.1110 99.1110 0.0000 99.1110 0.0000 99.1110 0.8890 0.0000 0.0000
42 99.1839 99.1839 0.0000 99.1839 0.0000 99.1839 0.8161 0.0000 0.0000
41 99.1997 99.1997 0.0000 99.1997 0.0000 99.1997 0.8003 0.0000 0.0000
40 99.2043 99.2043 0.0000 99.2043 0.0000 99.2043 0.7957 0.0000 0.0000
39 99.2068 99.2068 0.0000 99.2068 0.0000 99.2068 0.7932 0.0000 0.0000
38 99.2070 99.2070 0.0000 99.2070 0.0000 99.2070 0.7930 0.0000 0.0000
37 99.2072 99.2072 0.0000 99.2072 0.0000 99.2072 0.7928 0.0000 0.0000
35 99.2072 99.2072 0.0000 99.2072 0.0000 99.2072 0.7928 0.0000 0.0000
32 99.2073 99.2073 0.0000 99.2073 0.0000 99.2073 0.7927 0.0000 0.0000
3 99.9650 99.2073 0.0000 99.2073 0.0000 99.2073 0.0350 0.0000 0.7578
0 99.9650 99.2073 0.0000 99.2073 0.0000 99.2073 0.0350 0.0000 0.7578
•
written
by
mschmid
1 answer
Sorry about that! The problem is the trailing " /1" on the read names. Bwa trims this off but BBMap doesn't by default. So you can either remove it by mapping with BBMap and adding the "trd" flag (which stands for "trim read description"), or using the "addslash=f" flag when generating the synth reads with RandomReads.
Oh, I fixed it so the problem won't happen any more once I upload the latest version. It was unintentional.
Log in to answer this question.
That's odd! I'll look into it and get back to you...