MEDIPS : Populating fragments from Paired End reads
1
0
Entering edit mode
@kuldeep5534-7497
Last seen 6.7 years ago

Hi Lukas,

I have a question about MeDIP R script.

regions = data.frame(chr=as.character(as.vector(regions$rname)), start=as.numeric(as.vector(regions$pos)), stop=as.numeric(as.vector(regions$pos)+as.vector(regions$qwidth)-1), strand=as.character(as.vector(regions$strand)), isize=as.numeric(as.vector(regions$isize)), stringsAsFactors=F)

plus=regions$strand=="+" regions[plus, "stop"] = regions[plus, "stop"] + regions[plus, "isize"] + extend <------ line 153 regions[!plus, "start"] = regions[!plus, "start"] + regions[!plus,"isize" ] - extend <------ line 154 regions[, "stop"]=regions[, "stop"]+shift regions[, "start"]=regions[, "start"]+shift Let's take 'extend' to be zero, then template/fragment should be from start of first mate to end of second mate. Insertion length is defined as distance between start of first mate and end of second mate. So it should have been regions[plus, "stop"] = regions[plus, "start"] + regions[plus, "isize"] | | | | | | end of second mate start of first mate insertion size Your implementation is adding extra length which is equal to read length of first mate. And I have same opinion for line 154 as well. Have it been done intentionally? Please! Correct me if my understanding is wrong. Image link given below will help to understand if still the question is not clear. https://cloud.githubusercontent.com/assets/1637382/6764850/66028a8e-cfeb-11e4-8cbc-2605d664acb3.jpg Thanks for any help. medips • 787 views ADD COMMENT 0 Entering edit mode Lukas Chavez ▴ 570 @lukas-chavez-5781 Last seen 3.8 years ago USA/La Jolla/UCSD Hi kuldeep5534, thank you for your message! I have modified the according paired-end sequencing data import function to avoid adding length of the reads: regions[plus, "stop"] = regions[plus, "start"] + regions[plus, "isize"] + extend regions[!plus, "start"] = regions[!plus, "stop"] + regions[!plus,"isize" ] - extend This should soon be available in dev v. 1.17.4. All the best, Lukas > On 21 Mar 2015, at 12:32, kuldeep5534 [bioc] <noreply@bioconductor.org> wrote: > > Activity on a post you are following on support.bioconductor.org <https: support.bioconductor.org=""/> > User kuldeep5534 <https: support.bioconductor.org="" u="" 7497=""/> wrote Question: MEDIPS : Populating fragments from Paired End reads <https: support.bioconductor.org="" p="" 65901=""/>: > > > > Hi Lukas, > > I have a question about MeDIP R script. > > MEDIPS/R/MEDIPS.readRegionsFile.R > > > regions = data.frame(chr=as.character(as.vector(regions$rname)), start=as.numeric(as.vector(regions$pos)), stop=as.numeric(as.vector(regions$pos)+as.vector(regions$qwidth)-1), strand=as.character(as.vector(regions$strand)), isize=as.numeric(as.vector(regions$isize)), stringsAsFactors=F) > > plus=regions$strand=="+" > > regions[plus, "stop"] = regions[plus, "stop"] + regions[plus, "isize"] + extend <------ line 153 > > regions[!plus, "start"] = regions[!plus, "start"] + regions[!plus,"isize" ] - extend <------ line 154 > > regions[, "stop"]=regions[, "stop"]+shift > > regions[, "start"]=regions[, "start"]+shift > > > > Let's take 'extend' to be zero, then template/fragment should be from start of first mate to end of second mate. > > Insertion length is defined as distance between start of first mate and end of second mate. So it should have been > > > regions[plus, "stop"] = regions[plus, "start"] + regions[plus, "isize"] > > | | | > > | | | > > end of second mate start of first mate insertion size > > > Your implementation is adding extra length which is equal to read length of first mate. > > And I have same opinion for line 154 as well. > > > Have it been done intentionally? > > Please! Correct me if my understanding is wrong. > > > Image link given below will help to understand if still the question is not clear. > > https://cloud.githubusercontent.com/assets/1637382/6764850/66028a8e-cfeb-11e4-8cbc-2605d664acb3.jpg <https: cloud.githubusercontent.com="" assets="" 1637382="" 6764850="" 66028a8e-cfeb-11e4-8cbc-2605d664acb3.jpg=""> > Thanks for any help. > > > You may reply via email or visit MEDIPS : Populating fragments from Paired End reads >