Add best alignment score to abpoa_res_t

[subwaystation]

Hi @yangao07,

happy to see that abPOA is outperforming https://github.com/rvaser/spoa!

As indicated in #2, we are now trying to replace SPOA with abPOA in https://github.com/ekg/smoothxg/blob/57a9d568aec44986b8293ac1f33c1b128b8a0d46/src/smooth.cpp#L60.

I am hanging at the following:
When aligning a sequence to the abPOA graph, I need to know if the actual sequence or its reverse complement results in a better alignment score. I figured that best_score would deliver this, but it is not part of the abpoa_res_t. Could you add this, please?
Or ideally, abPOA already gives us this best alignment. I scanned the code, but could not find anything indicating that.
So, do you take the actual sequence and its reverse complement and perform alignments for both? Or just for the default sequence? Because we are still puzzled about this peculiar sequence in DRB1-3123. There is the rev_cigar argument, but it is only initialized as 0 and therefore does not have any effect, as far as I can tell.

Thanks for any feedback.

Best,
Simon

CC @ekg

[yangao07]

Hi Simon,

abPOA right now does align the 'reverse complementary' sequence automatically.
As already requested by Erik in #2, I plan to add this feature recently.
Also, we sure can add the best_score to the abpoa_res_t.
The rev_cigar is simply the reversed cigar strings, nothing related to the reverse complementary.

Yan

[ekg]

Is there any way to determine which orientation has been used? That's all we need to know to integrate abPOA into smoothxg.

…

On Thu, Sep 24, 2020, 03:30 Yan Gao ***@***.***> wrote: Hi Simon, abPOA right now does align the 'reverse complementary' sequence automatically. As already requested by Erik in #2 <#2>, I plan to add this feature recently. Also we sure can add the best_score to the abpoa_res_t. The rev_cigar is simply the reversed cigar strings, nothing related to the reverse complementary. Yan — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#4 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABDQEOBWAQ3RWCESR7KHM3SHKOLPANCNFSM4RW7RYRA> .

[yangao07]

Just updated the abPOA repo, please try out the ambiguous strand mode with '-s' (set abpt->amb_strand as 1).
Right now, the abpoa_res_t has both best_score and is_rc, and is_rc denotes whether or not the best score is from the reverse complementary sequence.

[subwaystation]

Awesome! That was quick.

I just tried again @ekg's example from #2 with -s:

./bin/abpoa ~/git/HLA-zoo/seqs/DRB1-3123.fa -r 4 -s  > DRB1-3123.abpoa.gfa
odgi build -g DRB1-3123.abpoa.gfa -o - | odgi unchop -i - -o - | odgi sort -i - -o - >DRB1-3123.abpoa.og
odgi viz -i DRB1-3123.abpoa.og -b -w 20 -y 500 -o DRB1-3123.abpoa.og.viz.inv.png -z

Now we see what we expected.
And thanks for adding best_score and is_rc into abpoa_res_t. That will ease our work!