ingest: Run Nextclade as part of ingest

[joverlee521]

Follow up to #40

With the recent addition of the community H5 Nextclade datasets in nextstrain/nextclade_data#196, it should now be possible to run Nextclade as part of ingest to assign clades to the H5 sequences.

Maybe this can replace the current manual clade labeling process with clade-labeling scripts?

[joverlee521]

I'm starting with the community/moncla-lab/iav-h5/ha/all-clades Nextclade dataset since that should work across fauna and NCBI sequences. Tested manually on ingest-with-nextclade branch.

NCBI

nextstrain build \
    ingest \
        joined-ncbi/results/nextclade.tsv \
        --configfile build-configs/ncbi/defaults/config.yaml

Almost everything gets assigned to the expected 2.3.4.4b clade, except 3 sequences were assigned to the 0 clade:

• A/skunk/Utah/24-008032-004/2024
• A/cattle/Texas/24-009499-002/2024
• A/cattle/Texas/24-009308-001/2024

Fauna

nextstrain build \
    --envdir ../env.d/seasonal-flu/ \
    ingest \
        fauna/results/nextclade.tsv \
        --configfile build-configs/ncbi/defaults/config.yaml

Since this is all avian flu and not just H5, there's ~30% not assigned to any clade.

See detailed breakdown of counts

clade	count	percent
	13371	30.65
0	598	1.37
1	630	1.44
1.1	127	0.29
1.1.1	76	0.17
1.1.2	224	0.51
2.1.1	80	0.18
2.1.2	79	0.18
2.1.3	55	0.13
2.1.3.1	36	0.08
2.1.3.2	429	0.98
2.1.3.2a	129	0.30
2.1.3.2b	133	0.30
2.1.3.3	51	0.12
2.2	817	1.87
2.2.1	555	1.27
2.2.1.1	140	0.32
2.2.1.1a	117	0.27
2.2.1.2	645	1.48
2.2.2	184	0.42
2.2.2.1	72	0.17
2.3.1	21	0.05
2.3.2	100	0.23
2.3.2.1	217	0.50
2.3.2.1a	932	2.14
2.3.2.1b	92	0.21
2.3.2.1c	211	0.48
2.3.2.1d	96	0.22
2.3.2.1e	793	1.82
2.3.2.1f	559	1.28
2.3.2.1g	384	0.88
2.3.3	28	0.06
2.3.4	574	1.32
2.3.4.1	58	0.13
2.3.4.2	86	0.20
2.3.4.3	206	0.47
2.3.4.4	307	0.70
2.3.4.4a	246	0.56
2.3.4.4b	14625	33.53
2.3.4.4c	1268	2.91
2.3.4.4d	134	0.31
2.3.4.4e	718	1.65
2.3.4.4f	173	0.40
2.3.4.4g	221	0.51
2.3.4.4h	736	1.69
2.4	6	0.01
2.5	18	0.04
3	7	0.02
4	26	0.06
5	16	0.04
6	11	0.03
7	69	0.16
7.1	20	0.05
7.2	60	0.14
8	4	0.01
9	22	0.05
Am-nonGsGD	1263	2.90
EA-nonGsGD	769	1.76

---

I'm going to join with metadata tomorrow Thursday and cross check the clades with the existing clades from fauna.

[joverlee521]

Latest push to the ingest-with-nextclade branch now joins the metadata with the Nextclade output.

I did a brief look into the fauna side to compare Nextclade clades with the existing clades

Of the 43,642 records

• 21,714 had the exact same clade designations between nextclade clade and gisaid clade.
• 13,242 had no clade designation in both
• 5590 had Nextclade clade but no gisaid clade
• 3096 had discrepancies between Nextclade clade and gisaid clade