Using MAQ with Illumina HiSeq results

[Lilach]

Hi,
I want to try MAQ (for the first time) for analysis of Illumina HiSeq human whole exome results, and I have two questions:

1. Is it ok to use hg19_chromFa.tar (from UCSC) as a reference, or should I run it again each chromosome individually ?

2. The text file I got from the HiSeq (which is fastq actually) is ok as an input for maq fasta2bfa command, or should I change its fastq format? I think that Hiseq output is already in the Sanger fastq format, but I'm not sure? below I copied the beginning of the txt file.

Thanks!



@ILLUMINA-FFC6C4_0005:7:1:1941:1087#0/1
CACATTGGATTGATCGGTCTCATTGGCCCCCCGGGAGAAGCTGGGGAGAAAGGAGATCAGGGGGTGCCAGGCGT
+ILLUMINA-FFC6C4_0005:7:1:1941:1087#0/1
faf\f_ccfffffSedaRe\dcdYffffcggg`g^ae`ca^RbaJ_VWW_Z\\Na`d``]a`bGM[UYVa`]`B
@ILLUMINA-FFC6C4_0005:7:1:2045:1092#0/1
GTGTGAATTTCATTTCCACATAAATTTTCTGAGCTGCATCACGGGAGATCCAGTTTGTACGAAGCCAGTTGTTT
+ILLUMINA-FFC6C4_0005:7:1:2045:1092#0/1
fffff_ffff[a\feaacafffffgfggcff]f]ae`beffadffafcf^[ac^dWd^abe`[d`_be_BBBBB
@ILLUMINA-FFC6C4_0005:7:1:2497:1094#0/1
GACGCTCACTCTCTCTGGTATAACTTCACCATCATTCATTTGCCCAGACATGGGCAACAGTGGTGTGAGGTCCA
+ILLUMINA-FFC6C4_0005:7:1:2497:1094#0/1
cbbd][dbdffffcbcccRa^ab^cff[d^Qaa^f[fefg_gcec`_f[Y^a^_ccaa_aZ`dYa[`Y``Y^Z[
@ILLUMINA-FFC6C4_0005:7:1:2730:1091#0/1
GGAACACACAGCTTCCCAGCTTTGGACAGTTGGTACAGCCTGAGGATGAGGGAAGCCAAGAACAAAAAACACCA
+ILLUMINA-FFC6C4_0005:7:1:2730:1091#0/1
ddaa`da`aacZa^BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB

[maubp]

I think that Hiseq output is already in the Sanger fastq format, but I'm not sure?

That string of BBBBBBBBB qualities in the fourth read looks like the PHRED Q2 marker for bad signal, which means this is in the old Illumina encoding, not the Sanger encoding used in the more recent Illumina pipelines. See:

http://news.open-bio.org/news/2010/04/illumina-q2-trim-fastq/

http://seqanswers.com/forums/showpost.php?p=17491&postcount=3

[dariober]

Hi,
I want to try MAQ (for the first time) for analysis of Illumina HiSeq human whole exome results, and I have two questions:

Do you have any particular reason to use MAQ? I think it is a bit outdated and nowdays most people use bwa, bowtie (1 or 2) or similar. I think MAQ was designed when sequence files were a few million reads and with the output of HiSeq (100s millions) it might take ages and/or require a lot of memory.

1. Is it ok to use hg19_chromFa.tar (from UCSC) as a reference, or should I run it again each chromosome individually ?

Not sure, but in general you don't want to split the reference sequence otherwise you can't tell whether a read aligns equally well to different chromosomes (well, you can but it would be more work downstream of the alignment which I don't think it pays off). To save time and parallelize you can split the sequence files though, unless it is RNAseq data you have.

Hope I'm not misunderstanding your question...

Best
Dario

[swbarnes2]

MAQ's problem is that it's not a fast aligner. You want one of the Burrows-Wheeler Transform algorithms. That means Bowtie or bwa.

And yes, in general, you want to align to the whole reference at one go. These algorithms will always try to fit your reads to the reference they are given, so you want to give the program your whole references. If a read aligns to Chr 6 perfectly, you don't want the software to be telling you it aligns to Chr 1 with two errors, but if you only give the software Chr 1 to align to, that's what it will do.

[Lilach]

Thank you for the answers!
So I read a little and it seems as Illimuna 1.5 fastq, becuase of the BBBBBBB strings.
Can I use BWA aln and sampe directly on these files, or should I reformat them to Sanger fastq?

Regarding MAQ - I wanted to compare its results to BWA. I already used BWA, but now I'm afraid the fastq qualities were not interpreted well?

[cam.jack]

I would have thought MAQ was totally outdated now. Does it even output to SAM?

With Illumina 1.3+ to 1.7 you need to use the -I flag with BWA.