What type of phred quality scores does samtools 1.3 expect

[gwilymh]

I am working with NG sequence data that was generated in 2012 using the Illumina HiSeq 1000. I have had some trouble getting samtools to accurately identify heterozygotes and phase bam files generated from these reads. I am wondering if said problems are a result of samtools misreading the phred scores in the fastq/bam files. Does anyone know the default method of phred-score encoding that samtools 1.3 expects?

e.g.
Sanger/Illumina 1.8 (ASCII 33 to 126)
Solexa/Illumina 1.0 (ASCII 59 to 126)
Illumina 1.3 (ASCII 64 to 126)

[dpryan]

All SAM/BAM/CRAM files expect Sanger/Illumina 1.8 format (aka, "phred + 33"). If you aligned phred+64 fastq files and didn't tell your aligner to fix the phred scores then you're going to have problems. I happen to have written a smaller conversion program if you happen to have done this (https://github.com/dpryan79/Answers/...iostars_133825).