BBMap and BBmerge insert sizs

Question

[Sharon]

Hi all

Using bbmerge, we have insert size 161 and using bbmap we have insert size 182. Here is my commands :

./bbmap.sh -Xmx24g in1=pair1.fastq.gz in2=pair2.fastq.gz ihist=hist_map.txt reads=400000 ref=genome.fa

./bbmerge.sh in1= in1=pair1.fastq.gz in2=pair2.fastq.gz ihist=hist_merge.txt reads=400000 prefilter=2 rem extend2=100

Which one I should go for? Also what is the count in hist.txt, the sum of counts is higher than the number of reads.

Answer 1

[GenoMax]

If you are able to map then the bbmap.sh size is more accurate since you have truth available via reference. With bbmerge command you are using you are doing some k-mer based extension of the reads, so keep that in mind. What do you get if you remove prefilter=2 rem extend2=100?

[Sharon]

If I still use reads=400000 and removed prefilter=2 rem extend2=100 I got 160 insert size.

[GenoMax]

Use the one from bbmap alignments.

[Sharon]

So if the insert size is 180, then 2 *150 reads are overlapping wih 120 right? This is still short insert size, right?

Insert size – 2 r = innermate 
180 -300 = -120 , right?

[GenoMax]

Yes. While the inserts are relatively short that is a characteristic of this library. It will have to be remade, if longer inserts are desired.

|------------------------------->30 bp|    R1 - 150 bp
|-------------------------------------|    Fragment 180 bp
|30 bp<-------------------------------|    R2 - 150 bp

[Sharon]

Thanks so much genomax