SAM to BAM --- samtools fail to read error

[shuoguo]

I am trying to transfer a SAM file to BAM file using samtools but get error as shown in the attached figure.

I have the SAM files ready in the right format, this SAM file is prepared by join another sam file with one bed file together. The columns are correct.

I used -S to indicate input is sam, and -b to indicate output is bam

i use -t to indicate that the sam file contains tab delimiter, and no header.

what could be the issue?

Thanks a lot

[shuoguo]

$ /usr/local/samtools-0.1.18/samtools view -bS all-hg18.bam -o test.bam
[samopen] no @SQ lines in the header.
[sam_read1] missing header? Abort!

$ /usr/local/samtools-0.1.18/samtools view -tbS all-hg18.bam -o test.bam
[main_samview] fail to open "all-hg18.bam" for reading.

[shuoguo]

find online manual said:

If your SAM file has header @SQ lines, you may get BAM by ...
If not, you need to have your reference file ref.fa and then do this:

looks like i need a ref.fa file, not sure what is that? how does it looks like and how to get it?

thanks

Shuoguo

[maubp]

That was shorthand for your reference genome in FASTA format.

[swbarnes2]

Try

samtools view -bSh mysam.sam > mybam.bam

I don't think you need the -t, and I'm not sure that the software likes you putting the -o option after the input file name.

-h will make sure the header goes on there.

[shuoguo]

It seems that no matter how I run it it will just spit out errors

$ /usr/local/samtools-0.1.18/samtools view -bS all-hg18-lifted.sam -o test.bam
[samopen] no @SQ lines in the header.
[sam_read1] missing header? Abort!

]$ /usr/local/samtools-0.1.18/samtools view -bS all-hg18-lifted.sam test.bam
[samopen] no @SQ lines in the header.
[main_samview] random alignment retrieval only works for indexed BAM files.

[shuoguo]

in my case i do not have teh original fasta files

That was shorthand for your reference genome in FASTA format.

[swbarnes2]

It seems that no matter how I run it it will just spit out errors

Use the caret to write to the .bam

/usr/local/samtools-0.1.18/samtools view -bS all-hg18-lifted.sam > test.bam

[shuoguo]

tried that with no luck. thanks for the reply.

Use the caret to write to the .bam

[maubp]

Which version of samtools do you have?

[shuoguo]

samtools-0.1.18

[dpryan]

Without knowing the chromosome names and sizes, you won't be able to make a BAM file (there'd be no way for it to know how to create the header). You can probably half-ass it by just getting a sorted list of chromosomes from your SAM file:

cat all-hg18-lifted.sam | cut -f 3 | sort | uniq > ref.fa.fai

For each line in ref.fa.fai, you then need to add that chromosome's length (separated by a tab from the chromosome name, so "chr1 500000000"). Obviously you don't know that, but you could just use a sufficiently large number. That would at least allow things to be converted to a BAM file. I don't know if that would screw things up with any downstream applications, but I can't currently think of a situation where it would.

[masterpiece]

As what dpryan and other said, its the header issue. May I know how you generate the sam file at the first place?

[shuoguo]

As what dpryan and other said, its the header issue. May I know how you generate the sam file at the first place?

Step 1: i have the original bam and bam.bai files downloaded from web. These are aligned with hg18.

Step 2: i transfer the same bam file to a bed file and a sam file.

Step 3: i liftover the bed file with the hg19 chain file (so the position is updated)

Step 4: I replace the position (3rd column) in the sam file with the liftover file.

Step 5: I want to transfer this sam file back to bam file, getting the error

Thanks for the help!!

[dpryan]

Ah, then since this is just hg18, you have a two options:

1) Do what I suggested above, substituting the length of the various chromosomes in hg18 for the made up values I originally suggested. You would have to look up these values.

2) Download the hg18 fasta file and run "samtools faidx" on it. You can then use the "-t" option that you tried before with the resulting .fai file. This is effectively the same as option 1, but requires less googling.

BTW, I hope you replaced all of the coordinate information rather than just the chromosome (position 3). Otherwise whatever you do downstream will be messed up.