Problem with creating GreenBlobList

[maikemuecke]

Hi, I have been using your package since roughtly December 2022 to check for unassigned Density between Protein Residues.
Since the release of v2.7.1 however I have been having issues with this. Density Peaks that are present in the ccp4 files (i check through coot) don't appear anymore in when I create the GreenBlobList and parse it's centroids.
One example where this problem occurs is the protein 3K9D. Between the sulfur (SG) of the cystein reside 359 & the nitrogen (NZ) of the Lysine residue 392 (both in chain A) is an apparent green density blob. With version 2.6.1 I was able to locate the blob through the following code:

analyzer=pdb_eda.densityAnalysis.fromPDBid('3k9d')
S_chain='A'
S_num=359
chain='A'
resnum=392
S_atom='SG'
N_atom='NZ'
S_coords=analyzer.biopdbObj[0][S_chain][int(S_num)][S_atom].get_coord()
print('S coords='+str(S_coords))
N_coords=analyzer.biopdbObj[0][chain][int(resnum)][N_atom].get_coord()
print('N coords='+str(N_coords))
greenBlobList = analyzer.greenBlobList
greenBlob_xyz=[i.coordCenter for i in greenBlobList]
greenBlob_xyz=np.vstack(greenBlob_xyz)
print(greenBlob_xyz.shape)
mid_point=(S_coords+N_coords)/2

if np.sum([np.linalg.norm(greenBlob_xyz-mid_point,axis=1) <=2])>0:
    idx_clossest_gree_blob=np.argmin(np.linalg.norm(greenBlob_xyz-mid_point,axis=1))
    print(greenBlobList[idx_clossest_gree_blob].volume)

In the code i simply check if within a 2 Angstrom radius of the midpoint between the SG & NZ atoms a centroid from the greenBlobList is located.
When I use version2.7.1 the greenBlobList does not contain the unassigned density peak at all. Generally the list does not contain any unassigned density peaks with negative xyz coordinates, even though part of the protein has negative structure coordinates.
I saw that from v2.6.1 to 2.7.1 the only change was the server from which the pdb files are downloaded, so I am very surprised at the issues I am facing. I was wondering if you could advise me what is going wrong.
Thank you!

[hunter-moseley]

Give me a few days. Working on a grant proposal submission. But I will look at this.

[hunter-moseley]

Are the comparisons between version results separated by time and/or being performed in separate directories?

The only rationale I have is that the files downloaded from wwPDB have changed.
The pdb_eda.densityAnalysis.fromPDBid function will pull down the files if they are not present in the ./ccp4_data and ./pdb_data subdirectories.

If there are separate subdirectories for producing the 2.6.1 and 2.7.1 results, you should be able to diff the ccp4 and .ent.gz files to see if they are different.

[maikemuecke]

Hi, yes they are seperated by time and performed in separate folders. Sadly I do not have the ccp4 file where pdb_eda.densityAnalysis.fromPDBid successfully identifies the unassigned density blob, as the screening was performed on our cluster (around March 2023). When I open the ccp4 file ith coot ( the one that is downloaded now when I run the screening), the density blob I am trying to identify is still there. However, it is not found by pdb_eda.densityAnalysis.fromPDBid.

…

________________________________ Von: Hunter Moseley ***@***.***> Gesendet: Samstag, 23. November 2024 21:05:36 An: MoseleyBioinformaticsLab/pdb_eda Cc: Mücke, Maike; Author Betreff: Re: [MoseleyBioinformaticsLab/pdb_eda] Problem with creating GreenBlobList (Issue #7) Are the comparisons between version results separated by time and/or being performed in separate directories? The only rationale I have is that the files downloaded from wwPDB have changed. The pdb_eda.densityAnalysis.fromPDBid function will pull down the files if they are not present in the ./ccp4_data and ./pdb_data subdirectories. — Reply to this email directly, view it on GitHub<#7 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ATRYSPISJXYQQXENYQONSU32CDNZBAVCNFSM6AAAAABSB6NFBWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIOJVGYZTSNRRHA>. You are receiving this because you authored the thread.Message ID: ***@***.***>