Bowtie 2 explanation of the pair-end mode report

[wariobrega]

Hello everybody. I'm a Master Student in Molecular Biology and Bioinformatics and pretty new to NGS technologies. I am using bowtie2 in pair-end mode after using bowtie1 in single end mode on some nanoCAGE RNA seq data. The reason I'm doing this is because bowtie1 in pair end mode has less customizable options than bowtie2 and is not possible to choose the size for the alignment window for the pair end mates, so I had very low alignment score against my reference genome (between 12 and 14% against my 90% and over in single end mode).

I cannot understand the report of the alignment score of Bowtie 2 in some parts. Here's an example:

16182999 reads; of these:
16182999 (100.00%) were paired; of these:
5731231 (35.42%) aligned concordantly 0 times
4522376 (27.95%) aligned concordantly exactly 1 time
5929392 (36.64%) aligned concordantly >1 times
----
5731231 pairs aligned concordantly 0 times; of these:
2381431 (41.55%) aligned discordantly 1 time
----
3349800 pairs aligned 0 times concordantly or discordantly; of these:
6699600 mates make up the pairs; of these:
3814736 (56.94%) aligned 0 times
1883429 (28.11%) aligned exactly 1 time
1001435 (14.95%) aligned >1 times
88.21% overall alignment rate

I have run bowtie 2 in pair end mode using the

-fr

option for the orientation of my reads.

according to the bolded part of the report, i have a lot of reads take in PE (because it specifies 6699600 mates out of 3349800 pairs which aligns nor concordantly nor discordantly (so they are not aligned on the genome). But still a fraction of these mates aligns. How is that possible? They shouldn't align at all!

also, I do not understand the overall score; Is there a place in which I can look to the criteria of this score? How does that 88.21% comes from?

Thanks for any kind of answer that you can give to me!

Daniele

[amitm]

I cannot understand the report of the alignment score of Bowtie 2 in some parts. Here's an example:

according to the bolded part of the report, i have a lot of reads take in PE (because it specifies 6699600 mates out of 3349800 pairs which aligns nor concordantly nor discordantly (so they are not aligned on the genome). But still a fraction of these mates aligns. How is that possible? They shouldn't align at all!

also, I do not understand the overall score; Is there a place in which I can look to the criteria of this score? How does that 88.21% comes from?

Thanks for any kind of answer that you can give to me!

Daniele

hi daniele,
the Bowtie2 result summary is divided in 3 sections:

• Concordant alignment - In your data (4522376 + 5929392) reads align concordantly. Which is 64.59% of reads

• Discordant alignment - So now 5731231 reads remain which is 35.41% (100-64.59). Of these, 2381431 reads align discordantly. That is to say, of the non-concordant fraction, 41.55% of reads (2381431 reads) align discordantly.

• The rest - Now, remember that alignment whether concord. or discord., but both are aligned in paired-end mode. The rest of the reads either align as singles (i.e. Read1 in one locus & Read2 in completely different locus or one mate aligned and the other unaligned) or may not align at all. So the reads that are in this section is Total -(Concord.+Discord.). That is 16182999 -(10451768+2381431) = 3349800 reads.
  Since alignment, if any, here is in single fashion so we calculate in mates (readsx2).

Now to reach the overall alignment, count the mates in total (i.e. mates aligned in paired and mates aligned in single fashion). That would be -
(10451768 x2)+(2381431 x2)+1883429+1001435 = 28551262 mates
That is 28551262 mates aligned of total (16182999 x2) mates, which is 88.21%.

[wariobrega]

hi daniele,
the Bowtie2 result summary is divided in 3 sections:

• Concordant alignment - In your data (4522376 + 5929392) reads align concordantly. Which is 64.59% of reads

• Discordant alignment - So now 5731231 reads remain which is 35.41% (100-64.59). Of these, 2381431 reads align discordantly. That is to say, of the non-concordant fraction, 41.55% of reads (2381431 reads) align discordantly.

• The rest - Now, remember that alignment whether concord. or discord., but both are aligned in paired-end mode. The rest of the reads either align as singles (i.e. Read1 in one locus & Read2 in completely different locus or one mate aligned and the other unaligned) or may not align at all. So the reads that are in this section is Total -(Concord.+Discord.). That is 16182999 -(10451768+2381431) = 3349800 reads.
  Since alignment, if any, here is in single fashion so we calculate in mates (readsx2).

Now to reach the overall alignment, count the mates in total (i.e. mates aligned in paired and mates aligned in single fashion). That would be -
(10451768 x2)+(2381431 x2)+1883429+1001435 = 28551262 mates
That is 28551262 mates aligned of total (16182999 x2) mates, which is 88.21%.

Thanks a lot for the explanation. So Bowtie 2 DOES align the reads in SE that do not align concordantly or discordantly in PE mode. But why this further alignment should be consider useful if it the reads does not match the PE criteria? Won't they probably align in other non significant regions and maybe false the output score?

Thanks a lot, also for the explanation of the Out score

[vittoria_1198]

Hi,
I am a very new user of bowtie. I used bowtie 2.0.0 to run a map step with single reads against a transcriptome reference.
At the end of the process, the alignment summary looed like this:

20000 reads; of these:
20000 (100.00%) were unpaired; of these:
1247 (6.24%) aligned 0 times
18739 (93.69%) aligned exactly 1 time
14 (0.07%) aligned >1 times
93.77% overall alignment rate

I am having hard time to understand the difference between the reads that align 1 time and the one that align more then 1 times. What does those last one mean? Should I considerate in my stats report only those theta align 1 time?
Moreover, I tried to take a look at the SAM file generated, and I could not be able to understand what is in it.
This is one line of the SAM file
@HD VN:1.0 SO:unsorted
@SQ SN:comp5504567_c0_seq1 LN:554

What are the meaning of each symbols?
Thanks so much for the help

Vittoria

[winsettz]

Hi,
I am a very new user of bowtie. I used bowtie 2.0.0 to run a map step with single reads against a transcriptome reference.
At the end of the process, the alignment summary looed like this:

20000 reads; of these:
20000 (100.00%) were unpaired; of these:
1247 (6.24%) aligned 0 times
18739 (93.69%) aligned exactly 1 time
14 (0.07%) aligned >1 times
93.77% overall alignment rate

I am having hard time to understand the difference between the reads that align 1 time and the one that align more then 1 times. What does those last one mean? Should I considerate in my stats report only those theta align 1 time?
Moreover, I tried to take a look at the SAM file generated, and I could not be able to understand what is in it.
This is one line of the SAM file
@HD VN:1.0 SO:unsorted
@SQ SN:comp5504567_c0_seq1 LN:554

What are the meaning of each symbols?
Thanks so much for the help

Vittoria

It suggests that some reads can theoretically align to more than one part of your reference.

For the SAM specification:

http://samtools.sourceforge.net/SAM1.pdf

Tag Description
@HD The header line. The rst line if present
VN* Format version. Accepted format: /^[0-9]+\.[0-9]+$/.
SO Sorting order of alignments. Valid values: unknown (default),unsorted, queryname andcoordinate.

@SQ Reference sequence dictionary. The order of @SQ lines defines the alignment sorting order
SN* Reference sequence name. Each @SQ line must have a unique SN tag
LN* Reference sequence length

[jajclement]

Hi!
Thank you a lot to amitm for his detailed answer. I had the same interpretation problem and it is now clearer for me after reading his reply to wariobrega.

I need another precision about the terms "concordantly" and "discondordantly": what does it mean exactly?

Thank you to all answers

[dpryan]

Hi!
Thank you a lot to amitm for his detailed answer. I had the same interpretation problem and it is now clearer for me after reading his reply to wariobrega.

I need another precision about the terms "concordantly" and "discondordantly": what does it mean exactly?

Thank you to all answers

Did you read the bowtie2 manual page? It's explained there.

[jajclement]

Thank you a lot for your answer.
In fact I just found the answer to my question...and was unable to clear my request.
Just a lesson to keep in mind: RTFM!!!

[chin]

Bowtie2

Hello, I am seriously new to NGS technologies, so there is a lot of non-understanding parts. I am currently new user for Bowtie2 therefore cannot understand the report of the alignment score of Bowtie2 results. Here's an example:
112291170 reads; of these:
95730389 (85.25%) were paired; of these:
78503591 (82.00%) aligned concordantly 0 times
1638154 (1.71%) aligned concordantly exactly 1 time
15588644 (16.28%) aligned concordantly >1 times
----
78503591 pairs aligned concordantly 0 times; of these:
1695209 (2.16%) aligned discordantly 1 time
----
76808382 pairs aligned 0 times concordantly or discordantly; of these:
153616764 mates make up the pairs; of these:
98441019 (64.08%) aligned 0 times
17302306 (11.26%) aligned exactly 1 time
37873439 (24.65%) aligned >1 times
16560781 (14.75%) were unpaired; of these:
4655378 (28.11%) aligned 0 times
1290964 (7.80%) aligned exactly 1 time
10614439 (64.09%) aligned >1 times
50.44% overall alignment rate
In previously, there was an useful explanation onto it but I still unable to get the exact amount for overall alignment rate-how does that 50.44% comes from?
Thanks for helping me.

[chin]

How about if there not 100% for aligned, such as:
112291170 reads; of these:
95730389 (85.25%) were paired; of these:
78503591 (82.00%) aligned concordantly 0 times
1638154 (1.71%) aligned concordantly exactly 1 time
15588644 (16.28%) aligned concordantly >1 times
----
78503591 pairs aligned concordantly 0 times; of these:
1695209 (2.16%) aligned discordantly 1 time
----
76808382 pairs aligned 0 times concordantly or discordantly; of these:
153616764 mates make up the pairs; of these:
98441019 (64.08%) aligned 0 times
17302306 (11.26%) aligned exactly 1 time
37873439 (24.65%) aligned >1 times
16560781 (14.75%) were unpaired; of these:
4655378 (28.11%) aligned 0 times
1290964 (7.80%) aligned exactly 1 time
10614439 (64.09%) aligned >1 times
50.44% overall alignment rate
How should I calculate its overall alignment rate?
Thank you.

[yksikaksi]

How about if there not 100% for aligned, such as:
112291170 reads; of these:
95730389 (85.25%) were paired; of these:
78503591 (82.00%) aligned concordantly 0 times
1638154 (1.71%) aligned concordantly exactly 1 time
15588644 (16.28%) aligned concordantly >1 times
----
78503591 pairs aligned concordantly 0 times; of these:
1695209 (2.16%) aligned discordantly 1 time
----
76808382 pairs aligned 0 times concordantly or discordantly; of these:
153616764 mates make up the pairs; of these:
98441019 (64.08%) aligned 0 times
17302306 (11.26%) aligned exactly 1 time
37873439 (24.65%) aligned >1 times
16560781 (14.75%) were unpaired; of these:
4655378 (28.11%) aligned 0 times
1290964 (7.80%) aligned exactly 1 time
10614439 (64.09%) aligned >1 times
50.44% overall alignment rate
How should I calculate its overall alignment rate?
Thank you.

Please give more information about the data above. A brief summary of your method and work will be hlepful for those who might able to help.

[shauryajauhari]

Using bowtie2, an alignment file (SAM) file was inferred. The question at hand is to report the count and sequences of multiple mapped reads.
As per the flags and tags specifications of the SAM format, I am trying to sift the alignment file on the basis of the following:
1. The value of N in each record carrying tags XS:i:<N> and AS:i:<N> must be identical.
2. The reads are preconditioned to be concordantly aligned, i.e. "YT:Z:CP" tag is set.
3. The flags 0x100 (256 bits) and 0x800 (2048 bits) represent secondary and supplementary mappings respectively.
Despite the aforementioned stipulations, the output is discrepant as compared to the alignment summary provided by bowtie2. Is there any alternative to the logic as stated.
Thanks in advance.
SJ.

[shauryajauhari]

How to derive multiple mapped reads from a SAM file?

MAPQ score of 255 engenders uniquely aligned reads; not to mention with the highest mapping score. Based on some literature, and by logic too, isn't it obvious that any number between 0 and 255 will represent multiple aligned reads. So, is it a subjective matter to chose quality score (depiction as per the 5th field in the SAM file) to ascertain multiple reads.
I believe the following will simply do the trick:
> samtools view -b -q <quality score> your_alignment_file.bam > filtered.bam
Additionally, could you also comment on the discrepancy (if any) with the alignment summary/ statistics from Bowtie2.

[gopal_botany]

Hello Everyone,

I have 2x250 paired-end data of bacteria, I have aligned reads to reference by bowtie2 and got the sam file output.

My query is, how to get the mapping position of read2 (mate read) as in sam alignment the mapping position remain same, See few lines of alignment

M01976:19:000000000-D240E:1:1101:14558:1508 73 16CFP21_Rv1984c 449 3 110M = 449 0 GTGGTTTCTCCAGCATGTTGTGGGTCGGCTTTTCGTTTCCGACAATCGTTCCGCTGTTTATCTCTAAGTCCATTAACTTGTGTGCT
CCCGACGTTCCATTTTGCACCGTT AFB0FFGEFAA1100B1DA2B////0/////012/>>/12/////B1//>00//>/?11212>>B2122211B2211>B1B1B01B11//////?0011?1?11110/0. AS:i:-45 XN:i:0 XM:i:15 XO:i:0 XG:i:0 NM:i:15 MD:Z:2
4G4G0G0G5G10G8A2G7A4A19A4A1A7G0A0 YT:Z:UP
M01976:19:000000000-D240E:1:1101:14558:1508 133 16CFP21_Rv1984c 449 0 * = 449 0 CCTTCCTGTTCGCCTCTATTTTCGCCGCCTGTCTTGTCATCCCCGTCTGTTCATTCGTTACATGCGCCTTATTATTTCCTCCTCCC
TTCCATTTTTGTTCTTCGTGTTCTC 21221011AA2/////112A2A2////////11@@1B12@21/////>/1222>2>200//1>21////?11B22>222110000/01001@22@1/00<1<1/0//0111 YT:Z:UP

My second query is how to find out the paired-read aligned in reverse order (strand) of the reference?

Please help.

Thanks.

Gopal