Question about sam output of Bismark

serenaliao · 02-12-2013, 09:54 AM

I am wondering if anyone has ever encountered similar problem. I used the latest version of Bismark for my data(pair end 100bp). After a closer look at the output sam file, I found there are some reads which is pretty short like this:
MWR-PRG-0014:95: D1HU8ACXX:6:1101:1417:2210_1:N:0:GCCAAT/1 115 chr5 75301330 255 100M = 75301318 -112 CAACCAAACAAAATACAACTATAAACTAAACTCTACCCATAAAACGCCTATTTACAAACTCCGTCCCCTAACCTAAAAAAACTAAAAAACCATTTCAAAC #B?@ACCBCDCB@AAADDC>>:
C@CCDA@:C:?>5BACCC=933CEBEEEEEE?HC=(A6DGGEEGIHEDBHCCFIIIIIGIHDC1EA4FHHDDD?D?;@ NM:i:23 XX:Z:2G14G2G1G1G2GG5G6GG6G3G7T7G4GGGG5GGGG10G2
XM:Z:..x..............x..x.h.h..xh.....x......hh..Z...x...h........X......x....xhhh.....xhhh..........x.. XR:Z:CT XG:Z:G
A
MWR-PRG-0014:95: D1HU8ACXX:6:1101:1417:2210_1:N:0:GCCAAT/2 179 chr5 75301318 255 100M = 75301330 -112 TCCCTCTCTCTACAACCAAACAAAATACAACTATAAACTAAACTCTACCCATAAAACGCCTATTTACAAACTCCGTCCCCTAACCTAAAAAAACTAAAAA @C@FFFDFHHFDHEEGIIGIJJ
JJJ<FGICFGJJIIGJIJJGJ@HIIIJJGGGIJEIBEHHIIIJIHEEEHHDFFFEACDD?CDDDDCD3@BDBDCC>CC NM:i:23 XX:Z:5T8G14G2G1G1G2GG5G6GG6G3G7T7G4GGGG5GGGG1-XM:Z:..............x..............x..x.h.h..xh.....x......hh..Z...x...h........X......x....xhhh.....xhhh. XR:Z:GA XG:Z:GA

which start at 75301318 and end at 75301330 at chromosome 5. (This is for Mmusculus mm9).
But if if get the exact sequence with mm9 by library BSgenome in R:

getSeq(Mmusculus, names= "chr5", start=75301318, end=75301330,as.character=TRUE)
[1] "TCCCTTTCTCTAC"

It seems it doesn't overlap with any bases in the read, then why is that map to the position.

Anyone knows the answer? I am curious to know. Thanks!

fkrueger · 02-12-2013, 11:25 AM

If I am not mistaken then the first mapping position in SAM format is the position where the read starts, and the second one is where the second read starts and not where it ends. Does that make sense?

serenaliao · 02-12-2013, 11:31 AM

Quote:

Originally Posted by fkrueger

If I am not mistaken then the first mapping position in SAM format is the position where the read starts, and the second one is where the second read starts and not where it ends. Does that make sense?

Ok, so in that case, the sequence between the two start position should still be mapped back to the sequence in the sam file right?

fkrueger · 02-12-2013, 11:37 AM

Yes it should, but looking at the 12bp you posted they do agree. You have to keep in mind that Sam format always displays the sequence as the forward stand and your read seems to align to the reverse stand.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Non-BS mismatch count in .SAM (esp. Bismark)?	mixter	Epigenetics	5	10-15-2012 05:57 AM
SAM tools question	jsun529	Bioinformatics	1	05-31-2011 08:13 AM
question about bwa sam file	peachgil	Bioinformatics	1	04-06-2011 09:53 AM
SAM tools output question	sowmyai	Bioinformatics	5	07-11-2010 06:38 PM
Bfast output and "Empty Sequence Dictionary" in .sam output	aiden	Bioinformatics	1	05-28-2010 07:50 PM

02-12-2013, 09:54 AM	#1
serenaliao Member Location: Pittsburgh Join Date: Jan 2013 Posts: 22	Question about sam output of Bismark I am wondering if anyone has ever encountered similar problem. I used the latest version of Bismark for my data(pair end 100bp). After a closer look at the output sam file, I found there are some reads which is pretty short like this: MWR-PRG-0014:95: D1HU8ACXX:6:1101:1417:2210_1:N:0:GCCAAT/1 115 chr5 75301330 255 100M = 75301318 -112 CAACCAAACAAAATACAACTATAAACTAAACTCTACCCATAAAACGCCTATTTACAAACTCCGTCCCCTAACCTAAAAAAACTAAAAAACCATTTCAAAC #B?@ACCBCDCB@AAADDC>>: C@CCDA@:C:?>5BACCC=933CEBEEEEEE?HC=(A6DGGEEGIHEDBHCCFIIIIIGIHDC1EA4FHHDDD?D?;@ NM:i:23 XX:Z:2G14G2G1G1G2GG5G6GG6G3G7T7G4GGGG5GGGG10G2 XM:Z:..x..............x..x.h.h..xh.....x......hh..Z...x...h........X......x....xhhh.....xhhh..........x.. XR:Z:CT XG:Z:G A MWR-PRG-0014:95: D1HU8ACXX:6:1101:1417:2210_1:N:0:GCCAAT/2 179 chr5 75301318 255 100M = 75301330 -112 TCCCTCTCTCTACAACCAAACAAAATACAACTATAAACTAAACTCTACCCATAAAACGCCTATTTACAAACTCCGTCCCCTAACCTAAAAAAACTAAAAA @C@FFFDFHHFDHEEGIIGIJJ JJJ<FGICFGJJIIGJIJJGJ@HIIIJJGGGIJEIBEHHIIIJIHEEEHHDFFFEACDD?CDDDDCD3@BDBDCC>CC NM:i:23 XX:Z:5T8G14G2G1G1G2GG5G6GG6G3G7T7G4GGGG5GGGG1-XM:Z:..............x..............x..x.h.h..xh.....x......hh..Z...x...h........X......x....xhhh.....xhhh. XR:Z:GA XG:Z:GA which start at 75301318 and end at 75301330 at chromosome 5. (This is for Mmusculus mm9). But if if get the exact sequence with mm9 by library BSgenome in R: getSeq(Mmusculus, names= "chr5", start=75301318, end=75301330,as.character=TRUE) [1] "TCCCTTTCTCTAC" It seems it doesn't overlap with any bases in the read, then why is that map to the position. Anyone knows the answer? I am curious to know. Thanks!

02-12-2013, 11:25 AM	#2
fkrueger Senior Member Location: Cambridge, UK Join Date: Sep 2009 Posts: 625	If I am not mistaken then the first mapping position in SAM format is the position where the read starts, and the second one is where the second read starts and not where it ends. Does that make sense?

02-12-2013, 11:37 AM	#4
fkrueger Senior Member Location: Cambridge, UK Join Date: Sep 2009 Posts: 625	Yes it should, but looking at the 12bp you posted they do agree. You have to keep in mind that Sam format always displays the sequence as the forward stand and your read seems to align to the reverse stand.