No query sequence in sam?

[skoren]

Thanks for the quick fix on the last issue. I am also seeing some mapped reads which don't have either the query sequence or qual, yet they seem to be mapped. I thought the query should always be supplied in the sam format in this case or am I wrong? Here is an example from the sam fastq.gz as in issue #3

b37c6230-9e91-4d7c-aa5d-853a2e6129b2_Basecall_Alignment_template        272     chr14_GL000009v2_random 65906   0       19S27M1D2M2D15M4I19M1D22M1D4M1D49M1I14M3D43M2I8M1D4M1D7M1D7M1D14M1D19M1I11M1D12M2D14M1D7M2D4M3D56M1D17M1I2M1D4M1D7M1D7M5D7M1D26M2D9M1D4M1D4M1D11M4D34M4D24M2D5M2D1M1I24M1D9M1D11M1D2M1D1M1D6M1D5M3D24M2D11M4D1M1D7M1D6M3D13M1I6M1I56M1D8M1D9M2D26M2D15M1D55M1D5M1D17M1D35M1D27M2D28M1D39M2D3M2D2M1D51M2I44M1D3M1D2M1D33M1I27M1D20M1D24M1D9M1D37M2D9M1D4M4D47M1D13M3D73M1D5M1D51M1D17M1D10M2D9M2D12M2D31M2D9M1D42M2I1M7D19M1D6M1D20M2D5M1D3M2D41M1I31M1D15M1D28M3D28M1D6M2D20M1D4M1D3M3I14M2D13M2D13M2D30M3D3M1D9M2D7M2D21M2D8M1D8M1D9M2D4M2D4M1D17M1D1M2D2M1D15M1D15M1I4M1D25M1I33M1D49M1D20M2D8M2D24M2I10M2D30M7D5M1D2M4D11M3D48M1D5M3D15M1I8M1D21M2D28M1D23M1I7M1D16M5D6M1D12M1D9M1D7M2D4M1D5M3D7M4D3M1D4M3D10M3D17M1I20M2D5M1D8M2D15M1I2M3D2M1D5M2D8M2D11M1D22M2D2M1D8M1D5M1D7M1D33M1D18M5D22M2D13M3D3M1D14M1D9M1D19M1D16M2D12M1D66M2D50M1D8M1D22M1D5M2D9M1D4M1D7M3D6M1D43M1D26M1D2M1D2M2I5M1D16M1I50M2D4M1D5M1D22M4D2M3D3M1D6M1D11M1D9M2D9M1D9M3D9M4D40M2D1M2D8M1D12M1D7M2D28M3D6M2D13M2D10M1D14M1D8M3D16M1I1M2D19M1D6M4D4M1D40M2D3M2D18M3D127M1I2M1D8M3D3M4D11M1D10M2D15M1D19M4D4M3D11M1D11M1D66M2I18M1D9M1I20M1D9M2D5M1I7M1I6M1I5M4D1M1I2M1I3M1I21M1D7M1D18M1I20M1D9M1D8M5D6M2D17M1D7M1D22M2D3M1D11M6D2M3D6M1I11M2D4M4D1M2D11M1D6M4D5M1I5M2D2M2D5M1I12M1I12M1D5M2D37M1D6M1I3M3D14M1D19M2D5M1D7M1D14M1D11M1D2M2D20M2D7M1D7M1I8M1D14M2D10M1D5M3D9M1D7M1D5M3D70M2D55M1D8M1I11M1D11M2D16M2D7M4D2M1I10M2D8M2D23M2D3M2D14M2I2M1D15M2D3M2D14M1D32M1D4M1D8M1D10M2I4M1D4M2D42M1D25M1I8M1D13M1D8M2D36M1D16M2D13M1D39M1I25M2D10M3D29M2D1M1D13M2D6M3D11M1D18M2D10M1D12M1D40M1D18M4I8M283I22M9I5M2D54M4D7M1I30M2I12M2D23M2D30M2D20M1D8M1D22M1D5M1D4M1D5M1D98M2D5M1D10M2D21M1D6M1D6M1D25M1D10M2I2M1D15M1D20M3D16M4D5M1I2M2I12M1D3M2D3M1D19M1D32M1D20M1D26M3D39M1D5M1D9M1D18M1I7M1D3M1D8M1D20M1D12M1D11M1D8M1D13M1D16M1D11M1D5M1D11M1D39M1D8M2D24M2D4M1D11M1I48M2I39M1D2M1D6M1D9M2I2M6D55M1D1M1D18M1D10M1D2M2D13M2D24M2D10M2D31M1D17M1D24M6D2M1D26M1D23M1D12M1D7M2D14M1I6M4D44M1D3M2D18M1D24M1I18M5D2M2D22M1D5M1D14M1D1M1D16M3D18M5D6M1I28M1D8M2D7M1D3M2D10M1D8M1D10M1D6M1D20M5D39M1D18M1D28M4I9M2D12M2D19M1D4M2D9M1D11M1D27M2D5M1D11M1D22M1D39M2D20M1D26M2D35M1D9M1D7M3D8M2D7M1D24M2D4M2D13M1D9M1D35M1D8M1D6M2D10M1D53M4D13M1D2M1D22M1D25M1D6M1D46M1I2M2D24M2D8M1D9M1D3M2D12M1D31M1D6M1D12M2D20M1D8M4D25M1D22M1D12M2D15M1D9M1D33M1D8M1D46M1D11M2D27M3D4M1D50M1D8M2D10M1D13M2D2M1I5M3D5M2D30M1I4M1D6M1D17M1D13M2D4M1D15M3D18M1D41M1D8M1D1M1D14M1D25M1D25M3D14M4D28M1D8M1D10M4D9M1D54M2D2M1D10M1D39M1D8M1D8M1I15M2D8M1D12M1D3M1I31M2D3M1D6M1D3M1D4M1D12M1I11M1D50M4D4M1D25M1D5M1D18M1D42M3D19M1D5M1D11M3D10M1D11M1D3M2D5M2D8M3I8M1D7M3D17M1D10M5D3M1D4M1D8M1I20M2D4M1D6M1I6M1D18M3D4M1D11M2D13M1D20M4D6M1D23M25S  *       0       0       *       *       cm:i:195        s1:i:2366       NM:i:1633       ms:i:9876       AS:i:10139      nn:i:0

[lh3]

See the flag "272", in particular the "256" bit. This is the so-called "secondary alignment", which usually don't have SEQ and QUAL to save disk space. BWA-MEM doesn't output secondary alignment by default. When developing minimap2, I found having secondary alignments are often useful for trouble shooting, so I made them outputted by default. Many tools, such as samtools pileup and depth, ignore secondary alignments.

If more people find this is confusing, I can switch off the feature by default.

[lh3]

Secondary alignments are basically alternative and often suboptimal alignments of a read. However, for long noisy reads, the best scoring alignment may not be the true alignment. That is why I thought outputting them may be useful.

[skoren]

Thanks for the info, I think it's fine to have the secondary alignments on. All the other aligners I had tried for these reads (minialign, bwa mem, ngm-lr, graph map) must have been defaulting to no secondary alignments as well which is why I hadn't seen that output before for this dataset.