Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Smoothxg error during poa alignment to subgraph #19

Closed
brettChapman opened this issue Oct 12, 2020 · 4 comments · Fixed by #20
Closed

Smoothxg error during poa alignment to subgraph #19

brettChapman opened this issue Oct 12, 2020 · 4 comments · Fixed by #20

Comments

@brettChapman
Copy link

Hi Erik

I tried running smoothxg on my GFA files (edyeet params = p99/a99/s50k, p95/a95/s50k, p90/a90/s50k, p90/a90/s100k), from chromosome 2H alignments across 4 different genomes.

I found all but the p99/a99/s50k alignment failed.

Heres the p99/a99/s50k output:

[path sgd sort]: 86.67% progress: iteration: 26, eta: 1.46, delta_max: 173155956.54, number of updates: 26956363
[path sgd sort]: 90.00% progress: iteration: 27, eta: 0.42, delta_max: 59465276.20, number of updates: 26956248
[path sgd sort]: 93.33% progress: iteration: 28, eta: 0.12, delta_max: 20829314.71, number of updates: 26957441
[path sgd sort]: 96.67% progress: iteration: 29, eta: 0.03, delta_max: 4379940.05, number of updates: 26957158
[path sgd sort]: 100.00% progress: iteration: 30, eta: 0.01, delta_max: 1955241.69, number of updates: 26956240
topological sort 26819945 of 26819945 ~ 100.0000%
[smoothxg::smoothable_blocks] computing blocks
[smoothxg::smoothable_blocks] computing blocks 100.00%%
[smoothxg::break_blocks] cutting blocks that contain sequences longer than max-poa-length (10000)
[smoothxg::break_blocks] cut 263466 blocks of which 2782 had repeats
[smoothxg::smooth_and_lace] applying abPOA (global alignment mode) to block 268333/268333 100.000%
[smoothxg::smooth_and_lace] sorting path_mappings
[smoothxg::smooth_and_lace] building final graph
[smoothxg::smooth_and_lace] adding graph 268333/268333 100.000%
[smoothxg::smooth_and_lace] embedding path fragment 290597/290597
[smoothxg::smooth_and_lace] verifying paths

The p95/a95/s50k output:

[path sgd sort]: 83.33% progress: iteration: 25, eta: 6.03, delta_max: 104674509.10, number of updates: 45525237
[path sgd sort]: 86.67% progress: iteration: 26, eta: 1.68, delta_max: 108659302.66, number of updates: 45523325
[path sgd sort]: 90.00% progress: iteration: 27, eta: 0.47, delta_max: 53257587.61, number of updates: 45529831
[path sgd sort]: 93.33% progress: iteration: 28, eta: 0.13, delta_max: 13299096.00, number of updates: 45531472
[path sgd sort]: 96.67% progress: iteration: 29, eta: 0.04, delta_max: 4763521.38, number of updates: 45531132
[path sgd sort]: 100.00% progress: iteration: 30, eta: 0.01, delta_max: 1273493.52, number of updates: 45524400
topological sort 27368108 of 27368108 ~ 100.0000%
[smoothxg::smoothable_blocks] computing blocks
[smoothxg::smoothable_blocks] computing blocks 100.00%%
[smoothxg::break_blocks] cutting blocks that contain sequences longer than max-poa-length (10000)
[smoothxg::break_blocks] cut 142012 blocks of which 1789 had repeats
[simd_abpoa_align_sequence_to_subgraph1] Error in cg_backtrack. (4)to block 157649/409308 38.516%
[smoothxg::smooth_and_lace] applying srun: error: node-20: task 0: Exited with exit code 138.516%

The p90/a90/s50k output:

[path sgd sort]: 86.67% progress: iteration: 26, eta: 2.35, delta_max: 71882224.52, number of updates: 148980422
[path sgd sort]: 90.00% progress: iteration: 27, eta: 0.60, delta_max: 33787250.39, number of updates: 148982175
[path sgd sort]: 93.33% progress: iteration: 28, eta: 0.15, delta_max: 7108062.96, number of updates: 148978394
[path sgd sort]: 96.67% progress: iteration: 29, eta: 0.04, delta_max: 2098275.19, number of updates: 148980490
[path sgd sort]: 100.00% progress: iteration: 30, eta: 0.01, delta_max: 1218520.58, number of updates: 148979518
topological sort 41166023 of 41166023 ~ 100.0000%
[smoothxg::smoothable_blocks] computing blocks
[smoothxg::smoothable_blocks] computing blocks 100.00%%
[smoothxg::break_blocks] cutting blocks that contain sequences longer than max-poa-length (10000)
[smoothxg::break_blocks] cut 57103 blocks of which 1000 had repeats
[simd_abpoa_align_sequence_to_subgraph1] Error in cg_backtrack. (4)to block 974/858858 0.113%
srun: error: node-2: task 0: Exited with exit code 1lignment mode) to block 1014/858858 0.118%

The p90/a90/s100k output:

[path sgd sort]: 86.67% progress: iteration: 26, eta: 2.34, delta_max: 45389367.41, number of updates: 145013869
[path sgd sort]: 90.00% progress: iteration: 27, eta: 0.60, delta_max: 17867948.24, number of updates: 145014736
[path sgd sort]: 93.33% progress: iteration: 28, eta: 0.15, delta_max: 8877562.74, number of updates: 145014706
[path sgd sort]: 96.67% progress: iteration: 29, eta: 0.04, delta_max: 2934047.45, number of updates: 145012933
[path sgd sort]: 100.00% progress: iteration: 30, eta: 0.01, delta_max: 405494.10, number of updates: 145014858
topological sort 40181737 of 40181737 ~ 100.0000%
[smoothxg::smoothable_blocks] computing blocks
[smoothxg::smoothable_blocks] computing blocks 100.00%%
[smoothxg::break_blocks] cutting blocks that contain sequences longer than max-poa-length (10000)
[smoothxg::break_blocks] cut 73145 blocks of which 1151 had repeats
[simd_abpoa_align_sequence_to_subgraph1] Error in cg_backtrack. (4)to block 14442/806980 1.790%
srun: error: node-19: task 0: Exited with exit code 1ignment mode) to block 14470/806980 1.793%

I noticed last week you added a new parameter -l to specify the max-poa-length, so I updated to the latest version of Smoothxg on Friday (9th October).

Even though setting at a99/p99 didn't provide any problems for smoothxg, I think a90/p90 is where the alignments will work well. I compared gene level alignments using a99/p99 with a90/p90 and a95/p95 using edyeet and wfmash, and it looks like anything above 90% starts to filter too many sequences out, leaving a fragmented alignment across the gene region.

@ekg
Copy link
Collaborator

ekg commented Oct 12, 2020 via email

@brettChapman
Copy link
Author

Thanks. Are the abPOA parameters for the b parameter embedded in the code or could I try smoothxg with the -Z parameter, which would use an different alignment mode?

@brettChapman
Copy link
Author

Or use the -S parameter with smoothxg?

@ekg
Copy link
Collaborator

ekg commented Oct 12, 2020 via email

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants