Question

ShortRead: how to convert qualities?

0

Entering edit mode

robmaz77 ▴ 20

@dc2bd3d5

Last seen 3.7 years ago

Is there a recommended way of converting FASTQ qualities? E.g., I read a phred64-encoded file with FastqStreamer, which gives me a ShortReadQ object with a quality of type SFastqQuality. After some manipulation I want to write it as phred33. For some reason I had thought I could write like this

> writeFastq(fq,file="foo.fq.gz",qualityType="FastqQuality")

No idea where I got that from, but it does not seem to work anyway.

I can come up with something like this

> fq2 <- ShortReadQ(sread=sread(fq),
                    id=id(fq),
                    quality=relist(BString(intToUtf8(t(as(quality(fq),"matrix"))+33L)),
                                   quality(fq)@quality))

# and in fact
> identical(as(quality(fq),"matrix"),as(quality(fq2),"matrix"))
[1] TRUE
> writeFastq(fq2,file="bar.fq.gz")

but this is hardly the pinnacle of elegance?

ShortRead FastqQuality • 1.7k views

ADD COMMENT • link 3.8 years ago robmaz77 ▴ 20

score 0 · Answer 1 · 2021-02-17

0

Entering edit mode

Martin Morgan 25k

@martin-morgan-1513

Last seen 4 months ago

United States

I think you can as(quality(fq), "PhredQuality").

ADD COMMENT • link 3.8 years ago Martin Morgan 25k

0

Entering edit mode

Hm, that seems to force the encoding on the existing BStringSet, but not re-encode the score:

> range(as(quality(fq),"matrix"))
[1]  2 41
> range(as(as(quality(fq),"PhredQuality"),"matrix"))
[1] 33 72

It also appears to be a different class:

> quality(fq)
class: SFastqQuality
quality:
BStringSet object of length 1000:
...
> as(quality(fq),"PhredQuality")
PhredQuality object of length 1000:
...

Apparently it can be used instead of a BstringSet in the constructor,

> fq2 <- ShortReadQ(id=id(fq),sread=sread(fq),quality=as(quality(fq),"PhredQuality"))
> fq2
class: ShortReadQ
length: 1000 reads; width: 100 cycles

but is detected again as solexa64

> quality(fq2)
class: SFastqQuality
quality:
PhredQuality object of length 1000:
...

regaining the original scores:

> range(as(quality(fq2),"matrix"))
[1]  2 41

ADD REPLY • link 3.8 years ago robmaz77 ▴ 20

0

Entering edit mode

Probably not worth wasting too much time on this, since the solution I had basically works.

For the record, here is a modified version that is slightly faster by avoiding the transpose and also does not rely on autodetection of the original format and rectangularity:

phred64To33 <- function(fq) {
  ShortReadQ(
    id=id(fq),
    sread=sread(fq),
    quality=FastqQuality(
      relist(
        as(intToUtf8(utf8ToInt(as.character(unlist(quality(quality(fq)))))-64L+33L),
             "BString"),
        quality(quality(fq)))))
}

ADD REPLY • link 3.8 years ago robmaz77 ▴ 20