Check the mapping quality of spike-in sequences

[liu-genomics]

possible problem

It seems there are many very short alignments, in the range of 4-6bps, in the alignment. I guess this is because in the alignment command, I removed alignments that hit more than twice in the genome. This should basically avoid the alignment of very short sequences when aligning to the whole genome. But when I used a very short sequence as a reference sequence, there is a great chance that a 4-mer or a 5-mer, would only have alignment at 1 position in the genome.

So in the clean up process, would eliminate those short sequences and also those that don't start from the core region (96bp到102bp on the reference sequence) of the spike-in sequence.

reference coordinates

I used the reverse of the sequence to do mapping, so the location of the true signal would be at position 92 on the reverse strand. So lulu's request from 96bp to 102 bp on the plus strand (wait, is Lulu talking about the earlier version of the spike-in? The new spike-in only has a length of 180bp, and the modification site is on the 89bp using the plus strand as reference).

OK. For the reference for PGC experiments, we should look at 89-95 mapping to the reverse strand. There are very few number of extremely short alignments in this range. But to be more accurate, also need to only count alignments that have "67M". The original read length should be 75bp.

samtools view -F 16 -m 67 \ 
../CHe-LHu-pl1-Lu9_S9_R1_001.171022_spike_in_reverse_complement.umi_encoded_adaptor_removed_no_mismatch.sorted.dedup.bam.sorted \
"171022_spikein_RC:89-95" > test4.sam

相关应急code放在 181028_MS_code，manuscript saving code.

#重点确认一下显示定量准确性的那个图还行不行