BFAST - Alignment for ABI or Illumina sequencing - with qualities

[nilshomer]

I wanted to let the community know about a new release of BFAST that now supports reads with quality scores.

http://genome.ucla.edu/bfast

It is designed to handle Illumina and ABI SOLiD data on the human whole-genome resequencing scale (billions of reads). It is multi-threaded and can be easily be parallelized on a cluster, or on your local desktop. You can easily tune it to handle large insertions or deletions (>10bp) in your alignment, SNPs, and even appropriately mapping ABI SOLiD color errors, all while maintaining speed and accuracy.

[valeu]

Dear Nils,

Can I run BFAST directly on GAII fastq data? They look like this:

@HWUSI-EAS454:5:1:0:149
ATTTCTCCACCTCCTCNCCCACCCCTTTTTTTTCCTTACTTCTTACTAAT
+HWUSI-EAS454:5:1:0:149
abaabaaaaab`baa]D\a]D]bbaaaZa][``aa_a]_ba_aa_aa``_
@HWUSI-EAS454:5:1:0:314
AGCCAGATCCTTACCCNCTCCACCTCTTTTTCTGTGTTTTATTTATGGTG
+HWUSI-EAS454:5:1:0:314
a_aaZS]Za__NDV_^DOYYMDZVYaY]Y___]ZWZZZ]QPNUNU^^QQR

Thanks in advance,

Valentina

[valeu]

Dear Nils,

While waiting for your answer, I run BFAST on a very small subset of my mate-pair data without changing the format. I took only 42 reads (21 mate-pairs).

These are Illumina data and they should be aligned to the genome like this:
<-(50bp)--(3000pb)--(50bp)->

This means that the left read should be align to the Crick strand and the right one to the Watson strand with a 3000bp spacer between them.

So I run "bfast match" and then "bfast localalign".

.bmf file was successfully created:

In total, found matches for 20 out of 21 reads.
************************************************************
Terminating successfully!
************************************************************

but localalign reported an error (parameters "-l 3000 -L 3"):

Performing alignment...
Currently on:
thread:1 [0]Assertion failed: 0 <= endRowStepOne && 0 <= endColStepOne, file AlignNTSpace.c, line 192
Abort (core dumped)

without "-L" I don't get an error:

Outputted alignments for 20 reads.
Outputted 1 reads for which there were no alignments.
Outputting complete.
************************************************************
Terminating successfully!
************************************************************

So my question is why I can get such an error, and if the parameter "-L 3" I use in localalign is what I need for my mate-pairs?

And also "-f" option, is it for mirroring or for fasta file?

Thank you in advance,

Valentina

[valeu]

Dear Nils,

Sorry to bother you again,

I have a question about the output format of BFAST (option "-O" in postprocess).

-O 1 gives almost all information I need but it is not a standard format.. So I would prefer to have output in gff or sam. But
-O 2 does not give me the information about chromosome
-O 3 does not give the information about strand...

And also it would be great to have the information about positions of mismatches and indels! Now I don't see how I can get it...

Thank you,

Valentina

[nilshomer]

Dear Nils,

Can I run BFAST directly on GAII fastq data? They look like this:

@HWUSI-EAS454:5:1:0:149
ATTTCTCCACCTCCTCNCCCACCCCTTTTTTTTCCTTACTTCTTACTAAT
+HWUSI-EAS454:5:1:0:149
abaabaaaaab`baa]D\a]D]bbaaaZa][``aa_a]_ba_aa_aa``_
@HWUSI-EAS454:5:1:0:314
AGCCAGATCCTTACCCNCTCCACCTCTTTTTCTGTGTTTTATTTATGGTG
+HWUSI-EAS454:5:1:0:314
a_aaZS]Za__NDV_^DOYYMDZVYaY]Y___]ZWZZZ]QPNUNU^^QQR

Thanks in advance,

Valentina

That should work, although the base qualities may not be scaled properly.

Dear Nils,

While waiting for your answer, I run BFAST on a very small subset of my mate-pair data without changing the format. I took only 42 reads (21 mate-pairs).

These are Illumina data and they should be aligned to the genome like this:
<-(50bp)--(3000pb)--(50bp)->

This means that the left read should be align to the Crick strand and the right one to the Watson strand with a 3000bp spacer between them.

So I run "bfast match" and then "bfast localalign".

.bmf file was successfully created:

In total, found matches for 20 out of 21 reads.
************************************************** **********
Terminating successfully!
************************************************** **********

but localalign reported an error (parameters "-l 3000 -L 3"):

Performing alignment...
Currently on:
thread:1 [0]Assertion failed: 0 <= endRowStepOne && 0 <= endColStepOne, file AlignNTSpace.c, line 192
Abort (core dumped)

without "-L" I don't get an error:

Outputted alignments for 20 reads.
Outputted 1 reads for which there were no alignments.
Outputting complete.
************************************************** **********
Terminating successfully!
************************************************** **********

So my question is why I can get such an error, and if the parameter "-L 3" I use in localalign is what I need for my mate-pairs?

And also "-f" option, is it for mirroring or for fasta file?

Thank you in advance,

Valentina

What version are you using (hopefully bfast.0.6.1c)? The "-f" option is for the fasta filename, the "-F" option is for mirroring (notice the case!).

Dear Nils,

Sorry to bother you again,

I have a question about the output format of BFAST (option "-O" in postprocess).

-O 1 gives almost all information I need but it is not a standard format.. So I would prefer to have output in gff or sam. But
-O 2 does not give me the information about chromosome
-O 3 does not give the information about strand...

And also it would be great to have the information about positions of mismatches and indels! Now I don't see how I can get it...

Thank you,

Valentina

The SAM format has information about strand. See the SAM spec. As for positions of mismatches and indels, use a variant caller or compare the alignment to the reference. It is implicit in the SAM format.

For more help, consider the BFAST help mailing list ([email protected]).