Tweet
Hi there,
I am trying to map single-end 33 bp SOLiD data using bwa. I mapped the data using bowtie (-v2 -m1 --best --strata i.e. uniquely mapping reads in the best strata of alignment allowing upto 2 mismatches) and got about 40% of reads mapping. I wanted to try bwa since it supports gapped alignment.
I have a few questions.
1. I have a very large number of reads (4,085,198,187) ... it’s nucleosome mapping data and I want the mapping to be conservative. Its ok if I lose reads but I want the mapping to be accurate. It’s the first time I am trying to use BWA since it is able to do gapped alignments, so I was interested in some suggestions regarding parameters. It seems like the default parameters for bwa aln are reasonably conservative. So do you suggest I just go with them?
The parameters I mainly want feedback on are
-n
-k
-o (max number of gaps)
-e (max gap extensions)
-R
-q
2. Also, I want to keep only uniquely mapping reads. It wasn’t clear to me at which stage this filtering occurs i.e. bwa aln or samse or postprocessing. Is there some parameter in bwa aln that controls this (similar to –m in bowtie)?
–R seems to be control the criterion for proceeding with suboptimal alignments so it doesn’t seem to be exactly what I am looking for?
Or is it the –n parameter in bwa samse?
Or do I have to postprocess the BAM output from bwa samse using the XT/X0 tag?
3. For the XT tag, what is the definition of unique. Is it that the number of hits in the best stratum of alignment must be 1 i.e. X0 tag must take value 1? Or does it mean that the total number of hits (including suboptimal ones) based on the bwa aln parameters must be 1?
4. I’m not clear what the –n INT parameter does in bwa samse : Does that only control what gets written in the XA tag? Or does it also filter out reads that map more than –n INT locations? What is the default value for –n INT? The manual says repetitive hits will be chosen randomly. So for a read that has more than 1 best hit, only one of the locations will be chosen at random and reported?
I apologize if these are naive questions. I did try to find answers online, by reading the manual and in the forum archives but didn’t get any information to my satisfaction.
Your input is greatly appreciated.
I am trying to map single-end 33 bp SOLiD data using bwa. I mapped the data using bowtie (-v2 -m1 --best --strata i.e. uniquely mapping reads in the best strata of alignment allowing upto 2 mismatches) and got about 40% of reads mapping. I wanted to try bwa since it supports gapped alignment.
I have a few questions.
1. I have a very large number of reads (4,085,198,187) ... it’s nucleosome mapping data and I want the mapping to be conservative. Its ok if I lose reads but I want the mapping to be accurate. It’s the first time I am trying to use BWA since it is able to do gapped alignments, so I was interested in some suggestions regarding parameters. It seems like the default parameters for bwa aln are reasonably conservative. So do you suggest I just go with them?
The parameters I mainly want feedback on are
-n
-k
-o (max number of gaps)
-e (max gap extensions)
-R
-q
2. Also, I want to keep only uniquely mapping reads. It wasn’t clear to me at which stage this filtering occurs i.e. bwa aln or samse or postprocessing. Is there some parameter in bwa aln that controls this (similar to –m in bowtie)?
–R seems to be control the criterion for proceeding with suboptimal alignments so it doesn’t seem to be exactly what I am looking for?
Or is it the –n parameter in bwa samse?
Or do I have to postprocess the BAM output from bwa samse using the XT/X0 tag?
3. For the XT tag, what is the definition of unique. Is it that the number of hits in the best stratum of alignment must be 1 i.e. X0 tag must take value 1? Or does it mean that the total number of hits (including suboptimal ones) based on the bwa aln parameters must be 1?
4. I’m not clear what the –n INT parameter does in bwa samse : Does that only control what gets written in the XA tag? Or does it also filter out reads that map more than –n INT locations? What is the default value for –n INT? The manual says repetitive hits will be chosen randomly. So for a read that has more than 1 best hit, only one of the locations will be chosen at random and reported?
I apologize if these are naive questions. I did try to find answers online, by reading the manual and in the forum archives but didn’t get any information to my satisfaction.
Your input is greatly appreciated.