Re-run coverage compute with new PIDs from release

[GaelleChapuis]

The Phys WG is interested to recompute the coverage per brain region using the new set of PIDs that are released (and that are thus used in the BW analysis).

• [ ] change the query bwm_query() so it uses the new release tag
• [ ] compute the coverage with the new query

This is not urgent, can be post-poned to January, but is quick to do if the query is settled.

New PIDs for second release: https://github.com/int-brain-lab/iblalyx/blob/main/releases/2023_Q4_IBL_et_al_BWM_2_eids_pids.csv Code: ibldevtools/Gaelle/PhysWG/coverage/scratch_45__BrainregionCoverage.py

[GaelleChapuis]

I am for now doing this merging to get the complete list of PIDs:

        ins_bw_df = bwm_query()
        ins_bw_df2 = pd.read_csv('/Users/gaelle/Documents/Git/int-brain-lab/iblalyx/releases/2023_Q4_IBL_et_al_BWM_2_eids_pids.csv')
        pid = np.unique(ins_bw_df['pid'].to_list() + ins_bw_df2['pid'].to_list())

(699 PIDs) I am trying to use the Ephys Atlas clusters table to do the cell count. Two issues come up :

• Warning: Not all BW PIDs are contained in the table (I have only 644)
• Some of the clusters have atlas_id set to NaN. Removing 10882 clusters. This is using the label ‘latest’

[GaelleChapuis]

With 644 PIDs, the values are:

Number of brain regions done: 75 too small and 66 (28.82096069868996%) needing more recording for overall 304. 163 brain regions done 71.17903930131004% brain region done.

---

Volume of brain region done: 7.409678101713824% for enough_rec label 0 91.52819726781084% for enough_rec label 1 1.0621246304753422% for enough_rec label too small

[GaelleChapuis]

Done with W51, 671 PIDs (for 699 total), 0 clusters with NaNs as acronyms, 91.8% coverage