Exporting Crosslinks to AlphaLink2


from pyXLMS import __version__
 
print(f"Installed pyXLMS version: {__version__}")

✓


    Installed pyXLMS version: 1.6.0


from pyXLMS import parser
from pyXLMS import exporter

All exporting functionality is available via the exporter submodule. We also import the parser submodule to read crosslinks.


parser_result = parser.read(
    "../../data/ms_annika/XLpeplib_Beveridge_QEx-HFX_DSS_R1_Crosslinks.xlsx",
    engine="MS Annika",
    crosslinker="DSS",
)
xls = parser_result["crosslinks"]

✓


    Reading MS Annika crosslinks...: 100%|██████████████████████████████████████████████████████████████████████████| 300/300 [00:00<00:00, 19063.00it/s]

We read crosslinks using the generic parser from a single MS Annika .xlsx file. For easier access we also assign our crosslinks to the variable xls.


from pyXLMS.transform import validate, targets_only
 
xls = targets_only(validate(xls))

✓


    Iterating over scores for FDR calculation...:  25%|██████████████████                                                       | 74/300 [00:00<?, ?it/s]

Important

For structure prediction with AlphaLink2 you would usually want to only consider high-confidence target crosslinks, either by pre-filtering in the search engine of your choice or via the pyXLMS functions validate() [docs ] and targets_only() [docs ]!


from pyXLMS.transform import filter_proteins
 
cas9 = filter_proteins(xls, proteins=["Cas9"])["Both"]

Important

You also should filter your crosslinks to only contain residue pairs of your protein(s) of interest, for example with the filter_proteins() [docs ] function!


result = exporter.to_alphalink2(
    cas9,
    fasta="../../data/_fasta/Cas9_plus10.fasta",
    annotated_fdr=0.01,
    filename_prefix="Cas9",
)

✓


    Exporting crosslinks to AlphaLink2...: 100%|███████████████████████████████████████████████████████████████████████████████| 223/223 [00:00<?, ?it/s]

The function exporter.to_alphalink2() exports a list of crosslinks to AlphaLink2 format for protein (complex) structure prediction with crosslink restraints. The tool AlphaLink2 is available via the link github.com/Rappsilber-Laboratory/AlphaLink2 . The function to_alphalink2() takes in a list of crosslinks as input and additionally needs a FASTA file as input, containing the sequences of the protein/chains of interest that should be modelled by AlphaLink2. Furthermore, the AlphaLink2 format requires an FDR column for your crosslinks. This is controlled via the annotated_fdr parameter which accepts a single float as input or a list of float values. In our case we specified annotated_fdr=0.01 which means that all crosslinks are labelled with a constant 1% false-discovery-rate (FDR). We could also pass a list of float values of equal length as our crosslinks to set individual FDR values for all crosslinks. You can read more about the to_alphalink2() function and all its parameters here: docs.

Important

If transform.annotate_fdr() [docs ] has been run previously, the to_alphalink2() function will automatically use the pyXLMS annotated FDR for the AlphaLink2 FDR column and only use the parameter annotated_fdr as a backup. To override this behaviour pass try_use_annotated_fdr=False to the to_alphalink2() function which will explicitly use the annotated_fdr parameter value(s)!


for key, value in result.items():
    print(f"{key}: {type(value)}")

✓


    AlphaLink2 crosslinks: <class 'str'>
    AlphaLink2 FASTA: <class 'str'>
    AlphaLink2 DataFrame: <class 'pandas.core.frame.DataFrame'>
    Exported files: <class 'list'>

The exporter function returns a dictionary

with key "AlphaLink2 crosslinks" containing the formatted crosslink input for AlphaLink2,
with key "AlphaLink2 FASTA" containing the FASTA file content for AlphaLink2,
with key "AlphaLink2 DataFrame" containing the exported crosslinks as a pandas DataFrame ,
and with key "Exported files" containing a list of filenames of all files that were written to disk.


print("\n".join(result["AlphaLink2 crosslinks"].split("\n")[:5]))

✓


    779 A 779 A 0.01
    866 A 866 A 0.01
    677 A 677 A 0.01
    677 A 48 A 0.01
    34 A 34 A 0.01

Via "AlphaLink2 crosslinks" we can access the formatted crosslink input for AlphaLink2, if filename_prefix was not None this has also been written to a file with name {filename_prefix}_AlphaLink2.txt.


print(result["AlphaLink2 FASTA"])

✓


    >A|Cas9
    GAASMDKKYSIGLAIGTNSVGWAVITDEYKVPSKKFKVLGNTDRHSIKKNLIGALLFDSGETAEATRLKRTARRRYTRRKNRICYLQEIFSNEMAKVDDSFFHRLEESFLVEEDKKHERHPIFGNIVDEVAYHEKYPTIYHLRKKLVDSTDKADLRLIYLALAHMIKFRGHFLIEGDLNPDNSDVDKLFIQLVQTYNQLFEENPINASGVDAKAILSARLSKSRRLENLIAQLPGEKKNGLFGNLIALSLGLTPNFKSNFDLAEDAKLQLSKDTYDDDLDNLLAQIGDQYADLFLAAKNLSDAILLSDILRVNTEITKAPLSASMIKRYDEHHQDLTLLKALVRQQLPEKYKEIFFDQSKNGYAGYIDGGASQEEFYKFIKPILEKMDGTEELLVKLNREDLLRKQRTFDNGSIPHQIHLGELHAILRRQEDFYPFLKDNREKIEKILTFRIPYYVGPLARGNSRFAWMTRKSEETITPWNFEEVVDKGASAQSFIERMTNFDKNLPNEKVLPKHSLLYEYFTVYNELTKVKYVTEGMRKPAFLSGEQKKAIVDLLFKTNRKVTVKQLKEDYFKKIECFDSVEISGVEDRFNASLGTYHDLLKIIKDKDFLDNEENEDILEDIVLTLTLFEDREMIEERLKTYAHLFDDKVMKQLKRRRYTGWGRLSRKLINGIRDKQSGKTILDFLKSDGFANRNFMQLIHDDSLTFKEDIQKAQVSGQGDSLHEHIANLAGSPAIKKGILQTVKVVDELVKVMGRHKPENIVIEMARENQTTQKGQKNSRERMKRIEEGIKELGSQILKEHPVENTQLQNEKLYLYYLQNGRDMYVDQELDINRLSDYDVDAIVPQSFLKDDSIDNKVLTRSDKNRGKSDNVPSEEVVKKMKNYWRQLLNAKLITQRKFDNLTKAERGGLSELDKAGFIKRQLVETRQITKHVAQILDSRMNTKYDENDKLIREVKVITLKSKLVSDFRKDFQFYKVREINNYHHAHDAYLNAVVGTALIKKYPKLESEFVYGDYKVYDVRKMIAKSEQEIGKATAKYFFYSNIMNFFKTEITLANGEIRKRPLIETNGETGEIVWDKGRDFATVRKVLSMPQVNIVKKTEVQTGGFSKESILPKRNSDKLIARKKDWDPKKYGGFDSPTVAYSVLVVAKVEKGKSKKLKSVKELLGITIMERSSFEKNPIDFLEAKGYKEVKKDLIIKLPKYSLFELENGRKRMLASAGELQKGNELALPSKYVNFLYLASHYEKLKGSPEDNEQKQLFVEQHKHYLDEIIEQISEFSKRVILADANLDKVLSAYNKHRDKPIREQAENIIHLFTLTNLGAPAAFKYFDTTIDRKQYRSTKEVLDATLIHQSITGLYETRIDLSQLGGD

Via "AlphaLink2 FASTA" we can access the FASTA file content for AlphaLink2, if filename_prefix was not None this has also been written to a file with name {filename_prefix}_AlphaLink2.fasta.


result["AlphaLink2 DataFrame"].head()

✓

png

Via "AlphaLink2 DataFrame" we can access the crosslink input for AlphaLink2 as a pandas DataFrame .


result["Exported files"]

✓


    ['Cas9_AlphaLink2.txt', 'Cas9_AlphaLink2.fasta']

Via "Exported files" we can access the names of all files that were written to disk.

Using the AlphaLink2 Exporter with `transform.annotate_fdr()`

In order to retrieve more accurate false-discovery-rate (FDR) estimates for the FDR column in the AlphaLink2 crosslink table we can use the pyXLMS function transform.annotate_fdr() [docs , page] which will calculate FDR estimates for all crosslinks.


parser_result = parser.read(
    "../../data/ms_annika/XLpeplib_Beveridge_QEx-HFX_DSS_R1_Crosslinks.xlsx",
    engine="MS Annika",
    crosslinker="DSS",
)
xls = parser_result["crosslinks"]

✓


    Reading MS Annika crosslinks...: 100%|██████████████████████████████████████████████████████████████████████████| 300/300 [00:00<00:00, 18977.32it/s]

We read crosslinks using the generic parser from a single MS Annika .xlsx file. For easier access we also assign our crosslinks to the variable xls.


from pyXLMS.transform import annotate_fdr
 
xls = annotate_fdr(xls)

✓


    Annotating FDR for crosslinks...: 100%|████████████████████████████████████████████████████████████████████████████████████| 300/300 [00:00<?, ?it/s]

Important

The function transform.annotate_fdr() [docs , page] needs to be run before any of the filtering steps in order to give accurate FDR estimates.


from pyXLMS.transform import validate, targets_only
 
xls = targets_only(validate(xls))

✓


    Iterating over scores for FDR calculation...:  25%|██████████████████                                                       | 74/300 [00:00<?, ?it/s]

Important


from pyXLMS.transform import filter_proteins
 
cas9 = filter_proteins(xls, proteins=["Cas9"])["Both"]

Important

You also should filter your crosslinks to only contain residue pairs of your protein(s) of interest, for example with the filter_proteins() [docs ] function!


result = exporter.to_alphalink2(
    cas9,
    fasta="../../data/_fasta/Cas9_plus10.fasta",
    annotated_fdr=0.01,
    try_use_annotated_fdr=True,
    filename_prefix=None,
)

✓


    Exporting crosslinks to AlphaLink2...: 100%|███████████████████████████████████████████████████████████████████████████████| 223/223 [00:00<?, ?it/s]

The function exporter.to_alphalink2() exports a list of crosslinks to AlphaLink2 format for protein (complex) structure prediction with crosslink restraints. The tool AlphaLink2 is available via the link github.com/Rappsilber-Laboratory/AlphaLink2 . The function to_alphalink2() takes in a list of crosslinks as input and additionally needs a FASTA file as input, containing the sequences of the protein/chains of interest that should be modelled by AlphaLink2. Furthermore, the AlphaLink2 format requires an FDR column for your crosslinks. This is controlled via the annotated_fdr parameter which accepts a single float as input or a list of float values. In our case we specified annotated_fdr=0.01 which means that all crosslinks are labelled with a constant 1% false-discovery-rate (FDR). We could also pass a list of float values of equal length as our crosslinks to set individual FDR values for all crosslinks. However, because we previously annotated the FDR for our crosslinks with transform.annotate_fdr() the value(s) passed via parameter annotated_fdr are only used as a backup, e.g. if the FDR calculated by transform.annotate_fdr() is nan. This behaviour is controlled by the parameter try_use_annotated_fdr=True which is by default True. You can read more about the to_alphalink2() function and all its parameters here: docs.

Tip

Specifying filename_prefix=None will only return the calculated results but not write them to disk!


result["AlphaLink2 DataFrame"].head()

✓

png

The FDR column now corresponds to the values that were calculated by transform.annotate_fdr(), even though we passed annotated_fdr=0.01.

Exporting Crosslinks to AlphaLink2

Using the AlphaLink2 Exporter with transform.annotate_fdr()

Using the AlphaLink2 Exporter with `transform.annotate_fdr()`