SAM to BAM conversion problem after HISAT2.

Question

[fmerkal]

Hello,

I recently used HISAT2 to align my paired-end reads to a reference genome. The command I used was hisat2 -q -p 4 -x path_to_index -1 reads_1.fastq -2 reads_2.fastq -S out.sam. This produced a SAM file that was ~14 GB and had headers that matched the ones I looked up online for this type of file.

I am having trouble converting my SAM output file to a BAM file using the following command samtools view -bS out.sam > out.bam. I should mention that the samtools version I am using is 1.4.1 and the syntax should be ok. The out.bam file is only 45 bytes and when I try to check it with samtools flagstat samtools flagstat out.bam, I get the following errors:

[W::sam_read1] parse error at line 1
[bam_flagstat_core] Truncated file? Continue anyway.

I have tried looking into this problem but couldn't find an answer that related specifically to my issue. Any help would be greatly appreciated. Cheers, Fjodor

[ATpoint]

Output if head out.sam and samtools view out.bam | head? Did you use something like nohup during alignment?

[fmerkal]

Hi, I did not use nohup. I am running hisat2 on a cluster and I received an output describing the alignment rate (overall alignment ~92%). I did not see other files, like the various .log files you would get if you ran Tophat. The output I get with head out.sam is listed below.

@HD     VN:1.0  SO:unsorted
@SQ     SN:MT   LN:16775
@SQ     SN:W    LN:6813114
@SQ     SN:Z    LN:82529921
@SQ     SN:1    LN:197608386
@SQ     SN:2    LN:149682049
@SQ     SN:3    LN:110838418
@SQ     SN:4    LN:91315245
@SQ     SN:5    LN:59809098
@SQ     SN:6    LN:36374701

When I run samtools view out.bam | head, I get

[W::sam_read1] parse error at line 1
[main_samview] truncated file.

[Istvan Albert]

note how above you run

samtools view out.bam

which is empty so will, of course, raise an error. See what

samtools view out.sam

prints

I think your BAM file is empty, for whatever reason, you would need to run the conversion again and see why it failed. I think it ran of out "something" and truncated the BAM file but the SAM file should be fine.

[eennadi]

You mean samtools sort -o out_sorted.bam out.sam worked for you? My problem is this generated .sam file which contains some alignments but when i ran samtools view -bS out.sam > out.bam both the out.sam and out.bam seem to corrupt the alignments such that nothing is both files

Curious to know how you went straight from samtools sort -o out_sorted.bam out.sam Did you sort bam first and then used sort first? Can you please be more detailed in your solution?

[ATpoint]

Please open a new question for this if you have trouble after using the search function and reading the manual. If opening a question, please make sure to include all commands you used.

Answer 1

[fmerkal]

Hello, I think I might have solved the problem. I used samtools sort to sort the SAM file and output it to BAM format with the command samtools sort -o out_sorted.bam out.sam. This seems to have created a BAM file that has a reasonable size ~2.9 GB (relative to the SAM file) and that I can view with samtools view out_sorted.bam.

This post was particularly helpful - htseq-count: error when reading beginning of SAM/BAM file. I still don't understand why it wasn't working when I used samtools view -bS out.sam > out.bam, but it seems fine for now. Thank you for all your help :)