Building a Glycopeptide Hypothesis from a FASTA File
- glycresoft build-hypothesis glycopeptide-fa
- Usage Example

Building a Glycopeptide Hypothesis from a FASTA File¶

The simplest way to build a glycopeptide database is to start with a list of theoretical glycoproteins in a FASTA file and perform a simple in-silico digest with one or more proteases, apply a set of modification rules, and combine the resulting peptides with a glycan hypothesis to produce glycopeptides.

Warning

Glycan compositions read from the associated glycan hypothesis must be classified as N-glycans, O-glycans, or GAG-Linkers in order for the algorithm to select the appropriate glycosite assignment algorithm. Glycan compositions lacking a classification will not be considered.

The build process is computationally intensive, as such this tool will spawn several worker processes to share the load.

glycresoft build-hypothesis glycopeptide-fa¶

Constructs a glycopeptide hypothesis from a FASTA file of proteins and a: collection of glycans.

glycresoft build-hypothesis glycopeptide-fa [OPTIONS] FASTA_FILE
                                            DATABASE_CONNECTION

Options

-z, --peptide-length-range <tuple>¶: The minimum and maximum peptide length to consider [default: 5, 60]

-g, --glycan-source <string>¶: The path, identity, or other specifier for the glycan source [required]

-s, --glycan-source-type <choice>¶: The type of glycan information source to use [default: text]

Choices: [

combinatorial; text; hypothesis;

analysis]

-G, --glycan-source-identifier <string>¶: The name or id number of the hypothesis or analysis to be used when using those glycan source types.

-p, --processes <int>¶: Number of worker processes to use. Defaults to 4 or the number of CPUs, whichever is lower [default: 4]

-n, --name <string>¶: The name for the hypothesis to be created

-u, --occupied-glycosites <int>¶: The number of occupied glycosylation sites permitted. [default: 1]

-e, --enzyme <string>¶: The proteolytic enzyme to use during digestion. May be specified multiple times, generating a co-digestion. May specify an enzyme name or a regular expression describing the cleavage pattern. Recognized enzyme names are: 2-iodobenzoate, alpha-lytic protease, arg-c, asp-n, asp-n_ambic, bnps-skatole, caspase 1, caspase 10, caspase 2, caspase 3, caspase 4, caspase 5, caspase 6, caspase 7, caspase 8, caspase 9, chymotrypsin, chymotrypsin high specificity, chymotrypsin low specificity, clostripain, cnbr, enterokinase, factor xa, formic acid, glu-c, glutamyl endopeptidase, granzyme b, hydroxylamine, iodosobenzoic acid, leukocyte elastase, lys-c, lys-c/p, none, ntcb, pepsin ph1.3, pepsin ph2.0, pepsina, proline endopeptidase, proteinase k, staphylococcal peptidase i, thermolysin, thrombin, trypchymo, trypsin, trypsin/p, v8-de, v8-e [default: trypsin] (May specify more than once)

-m, --missed-cleavages <int>¶: The number of missed proteolytic cleavage sites permitted [default: 1]

-c, --constant-modification <string>¶: Peptide modification rule which will be applied constantly (May specify more than once)

-v, --variable-modification <string>¶: Peptide modification rule which will be applied variablely (May specify more than once)

-V, --max-variable-modifications <int>¶: The maximum number of variable modifications that can be applied to a single peptide [default: 4]

-y, --semispecific-digest¶: Apply a semispecific enzyme digest permitting one peptide terminal to be non-specific [default: False]

-R, --reverse¶: Reverse protein sequences [default: False]

-F, --not-full-crossproduct¶: Do not produce full crossproduct. For when the search space is too large to enumerate, store, and load. [default: False]

--retain-all-peptides¶: Do not require a glycosylation site when saving base peptides [default: False]

-C, --include-n-x-c-glycosylation¶: Whether to support N-x-C in the N-glycosylation sequon [default: False]

-a, --annotation-path <path>¶: The path to an annotation XML file from UniProt, or ‘-’ to suppress annotations.

Arguments

FASTA_FILE¶: Required argument <path> A file containing protein sequences in FASTA format

DATABASE_CONNECTION¶: Required argument <string> A connection URI for a database, or a path on the file system

Usage Example¶

Below is a basic example of how to use this tool

# We'll build a simple glycopeptide hypothesis from just two glycoproteins, the
# isoforms of Alpha-1-acid glycoprotein.
$ cat agp.fa
>sp|P02763|A1AG1_HUMAN Alpha-1-acid glycoprotein 1 OS=Homo sapiens GN=ORM1 PE=1 SV=1
MALSWVLTVLSLLPLLEAQIPLCANLVPVPITNATLDQITGKWFYIASAFRNEEYNKSVQ
EIQATFFYFTPNKTEDTIFLREYQTRQDQCIYNTTYLNVQRENGTISRYVGGQEHFAHLL
ILRDTKTYMLAFDVNDEKNWGLSVYADKPETTKEQLGEFYEALDCLRIPKSDVVYTDWKK
DKCEPLEKQHEKERKQEEGES

>sp|P19652|A1AG2_HUMAN Alpha-1-acid glycoprotein 2 OS=Homo sapiens GN=ORM2 PE=1 SV=2
MALSWVLTVLSLLPLLEAQIPLCANLVPVPITNATLDRITGKWFYIASAFRNEEYNKSVQ
EIQATFFYFTPNKTEDTIFLREYQTRQNQCFYNSSYLNVQRENGTVSRYEGGREHVAHLL
FLRDTKTLMFGSYLDDEKNWGLSFYADKPETTKEQLGEFYEALDCLCIPRSDVMYTDWKK
DKCEPLEKQHEKERKQEEGES

# This hypothesis will include a combinatorial glycan composition hypothesis
# defined by these rules
$ cat combinatorial-rules.txt
Hex 3 12
HexNAc 2 10
Fuc 0 5
Neu5Ac 0 4

Fuc < HexNAc
HexNAc > NeuAc + 1

# We'll permit the following additional options:
#  -u 1 means that only one glycosylation site will ever be occupied
#  -e trypsin means that the proteins will be *in-silico* digested with trypsin
#  -m 1 means that we will only permit one missed cleavage by the protease
#  -c "Carbamidomethyl (C)" means we will have a constant modification "Carbamidomethyl"
#   on all Cysteine residues
#  -v "Deamidation (N)" means we will have a variable modification "Deamidation" on any
#   Asparagine residue
#  -v "Pyro-glu from Q (Q@N-term)" means that we will have a variable modification "Pyro-glu from Q"
#   on any Glutamine residue on the N-terminus of the peptide sequence
#  -p 4 means that this task will spawn four worker processes to share the work between
$ glycresoft build-hypothesis glycopeptide-fa -g combinatorial-rules.txt -s combinatorial\
    -u 1 -e trypsin -m 1 -c "Carbamidomethyl (C)" -v "Deamidation (N)"\
    -v "Pyro-glu from Q (Q@N-term)" -p 4 -n "Alpha-1-acid Glycopeptide Hypothesis"\
    agp.fa fasta-glycopeptides.db

2017-08-31 02:50:08.610441 Begin Combinatorial Glycan Hypothesis Serializer
{'derivatization': None,
 'engine': Engine(sqlite:///fasta-glycopeptides.db),
 'glycan_file': 'combinatorial-rules.txt',
 'loader': None,
 'reduction': None,
 'start_time': datetime.datetime(2017, 8, 31, 2, 50, 8, 608881),
 'status': 'started',
 'transformer': None,
 'uuid': '93604f0929794e2bab9df23e29118b09'}
2017-08-31 02:50:08.628096 Generating Glycan Compositions from Symbolic Rules for GlycanHypothesis(id=1, name=GlycanHypothesis-93604f0929794e2bab9df23e29118b09)
2017-08-31 02:50:15.188556 1000 glycan compositions created
2017-08-31 02:50:20.824145 Generated 1900 glycan compositions
2017-08-31 02:50:20.835196 Hypothesis Completed
2017-08-31 02:50:20.835577 End Combinatorial Glycan Hypothesis Serializer
2017-08-31 02:50:20.835661 Started at 2017-08-31 02:50:08.608881.
Ended at 2017-08-31 02:50:20.835265.
Total time elapsed: 0:00:12.226384
CombinatorialGlycanHypothesisSerializer completed successfully.
2017-08-31 02:50:20.903601 Begin Multiple Process Fasta Glycopeptide Hypothesis Serializer
{'constant_modifications': [Carbamidomethyl:57.021464],
 'engine': Engine(sqlite:///fasta-glycopeptides.db),
 'fasta_file': 'agp.fa',
 'max_glycosylation_events': 1,
 'max_missed_cleavages': 1,
 'n_processes': 4,
 'protease': 'trypsin',
 'start_time': datetime.datetime(2017, 8, 31, 2, 50, 20, 899793),
 'status': 'started',
 'total_glycan_combination_count': -1,
 'uuid': '1e6d202801ce4417b3b153ac84732401',
 'variable_modifications': [Deamidated:0.984016, Gln->pyro-Glu:-17.026549]}
2017-08-31 02:50:20.903709 Extracting Proteins
2017-08-31 02:50:20.927056 Digesting Proteins
2017-08-31 02:50:28.965977 205 Base Peptides Produced
2017-08-31 02:50:28.966163 Begin Applying Protein Annotations
2017-08-31 02:50:29.726805 ... Extracting Best Peptides
2017-08-31 02:50:29.753947 ... Building Mask
2017-08-31 02:50:29.755910 ... Removing Duplicates
2017-08-31 02:50:29.765368 ... Complete
2017-08-31 02:50:29.768705 Combinating Glycans
2017-08-31 02:50:30.766838 ... Building combinations for Hypothesis 1
2017-08-31 02:50:33.867166 1900 Glycan Combinations Constructed.
2017-08-31 02:50:33.867282 Building Glycopeptides
2017-08-31 02:50:33.895503 Begin Creation. Dropping Indices
2017-08-31 02:50:33.939597 ... Processing Glycan Combinations 0-1900 (100.00%)
2017-08-31 02:50:34.021121 ... Dealt Peptides 0-11 4.60%
2017-08-31 02:50:34.021361 ... Dealt Peptides 11-22 9.21%
2017-08-31 02:50:34.021429 ... Dealt Peptides 22-33 13.81%
2017-08-31 02:50:34.021495 ... Dealt Peptides 33-44 18.41%
2017-08-31 02:50:34.021560 ... Dealt Peptides 44-55 23.01%
2017-08-31 02:50:34.021624 ... Dealt Peptides 55-66 27.62%
2017-08-31 02:50:34.021781 ... Dealt Peptides 66-77 32.22%
2017-08-31 02:50:34.021890 ... Dealt Peptides 77-88 36.82%
2017-08-31 02:50:34.022035 ... Dealt Peptides 88-99 41.42%
2017-08-31 02:50:34.022238 ... Dealt Peptides 99-110 46.03%
2017-08-31 02:50:34.340675 ... Dealt Peptides 110-121 50.63%
2017-08-31 02:50:34.367891 ... Dealt Peptides 121-132 55.23%
2017-08-31 02:50:34.376397 ... Dealt Peptides 132-143 59.83%
2017-08-31 02:50:34.410928 ... Dealt Peptides 143-154 64.44%
2017-08-31 02:50:46.488753 ... Dealt Peptides 154-165 69.04%
2017-08-31 02:50:48.904966 ... Dealt Peptides 165-176 73.64%
2017-08-31 02:50:52.744931 ... Dealt Peptides 176-187 78.24%
2017-08-31 02:50:52.765699 ... Dealt Peptides 187-198 82.85%
2017-08-31 02:50:52.806290 ... Dealt Peptides 198-209 87.45%
2017-08-31 02:50:59.408814 ... Dealt Peptides 209-220 92.05%
2017-08-31 02:51:08.196088 ... Dealt Peptides 220-231 96.65%
2017-08-31 02:51:11.241885 ... Dealt Peptides 231-239 100.00%
2017-08-31 02:51:11.241991 ... All Peptides Dealt
2017-08-31 02:51:33.927773 ... 130001 Glycopeptides Created
2017-08-31 02:52:10.984681 Process 3560 completed. (41 peptides, 57000 glycopeptides)
2017-08-31 02:52:11.928619 Process 3556 completed. (66 peptides, 55100 glycopeptides)
2017-08-31 02:52:12.924645 Process 3558 completed. (66 peptides, 57000 glycopeptides)
2017-08-31 02:52:17.504664 Process 3562 completed. (66 peptides, 62700 glycopeptides)
2017-08-31 02:52:19.507377 Joining Process 3556 (False)
2017-08-31 02:52:19.507669 Joining Process 3558 (False)
2017-08-31 02:52:19.507839 Joining Process 3560 (False)
2017-08-31 02:52:19.507991 Joining Process 3562 (False)
2017-08-31 02:52:19.508170 All Work Done. Rebuilding Indices
2017-08-31 02:52:21.342381 Analyzing Indices
2017-08-31 02:52:21.605914 Done Analyzing Indices
2017-08-31 02:52:21.645477 Generated 231800 glycopeptides
2017-08-31 02:52:21.645579 Done
2017-08-31 02:52:21.659734 Hypothesis Completed
2017-08-31 02:52:21.660019 End Multiple Process Fasta Glycopeptide Hypothesis Serializer
2017-08-31 02:52:21.660143 Started at 2017-08-31 02:50:20.899793.
Ended at 2017-08-31 02:52:21.659825.
Total time elapsed: 0:02:00.760032
MultipleProcessFastaGlycopeptideHypothesisSerializer completed successfully.

If you instead wish to use an existing glycan hypothesis instead of creating a new one, you can modify the instructions above:

$ glycresoft build-hypothesis glycan-combinatorial rules-file.txt combinatorial-database -n "Combinatorial Human N-Glycans"
...

$ glycresoft build-hypothesis glycopeptide-fa -g combinatorial-database.db -s hypothesis -G 1\
    -u 1 -e trypsin -m 1 -c "Carbamidomethyl (C)" -v "Deamidation (N)"\
    -v "Pyro-glu from Q (Q@N-term)" -p 4 -n "Alpha-1-acid Glycopeptide Hypothesis"\
    agp.fa fasta-glycopeptides.db
...

The primary difference here is that the value of the -g option is the path to the source database’s path (or connection URI), -s indicates that the source is a hypothesis, and the new -G option identifies the hypothesis to use, as a single database may contain many hypotheses. The value of -G can be the ID of the hypothesis or the hypothesis’s name (in this case “Combinatorial Human N-Glycans”).

Supported Proteases¶

Enzyme Name	Recognized Pattern
bnps-skatole	W
caspase 1	(?<=[FWYL]w[HAT])D(?=[^PEDQKR])
caspase 10	(?<=IEA)D
caspase 2	(?<=DVA)D(?=[^PEDQKR])
caspase 3	(?<=DMQ)D(?=[^PEDQKR])
caspase 4	(?<=LEV)D(?=[^PEDQKR])
caspase 5	(?<=[LW]EH)D
caspase 6	(?<=VE[HI])D(?=[^PEDQKR])
caspase 7	(?<=DEV)D(?=[^PEDQKR])
caspase 8	(?<=[IL]ET)D(?=[^PEDQKR])
caspase 9	(?<=LEH)D
chymotrypsin	(?<=[FYWL])(?!P)
chymotrypsin high specificity	([FY](?=[^P]))\|(W(?=[^MP]))
chymotrypsin low specificity	([FLY](?=[^P]))\|(W(?=[^MP]))\|(M(?=[^PY]))\|(H(?=[^DMPW]))
clostripain	R
cnbr	M
enterokinase	(?<=[DE]{3})K
factor xa	(?<=[AFGILTVM][DE]G)R
formic acid	D
glutamyl endopeptidase	E
glu-c	E
granzyme b	(?<=IEP)D
hydroxylamine	N(?=G)
iodosobenzoic acid	W
lys-c	(?<=K)(?!P)
ntcb	w(?=C)
pepsin ph1.3	((?<=[^HKR][^P])[^R](?=[FLWY][^P]))\|((?<=[^HKR][^P])[FLWY](?=w[^P]))
pepsin ph2.0	((?<=[^HKR][^P])[^R](?=[FL][^P]))\|((?<=[^HKR][^P])[FL](?=w[^P]))
proline endopeptidase	(?<=[HKR])P(?=[^P])
proteinase k	[AEFILTVWY]
staphylococcal peptidase i	(?<=[^E])E
thermolysin	[^DE](?=[AFILMV])
thrombin	((?<=G)R(?=G))\|((?<=[AFGILTVM][AFGILTVWA]P)R(?=[^DE][^DE]))
trypsin	([KR](?=[^P]))\|((?<=W)K(?=P))\|((?<=M)R(?=P))
none	^&$
trypchymo	(?<=[FYWLKR])(?!P)
trypsin/p	(?<=[KR])
lys-c/p	(?<=K)
pepsina	(?<=[FL])
v8-de	(?<=[BDEZ])(?!P)
v8-e	(?<=[EZ])(?!P)
2-iodobenzoate	(?<=W)
arg-c	(?<=R)(?!P)
asp-n_ambic	(?=[DE])
asp-n	(?=[BD])
leukocyte elastase	(?<=[ALIV])(?!P)
alpha-lytic protease	[TASV]

Supported Modification Rules¶

glycresoft supports the full range of UNIMOD modification rules as well as some common alternative namings.

To be more specific, you are able to override modification targets when you specify modification names by passing the permitted target rules enclosed in parentheses following the modification name. For example “Deamidation (N)” will only target Asparagine residues, unlike the plain “Deamidation” rule which will target both Asparagine and Glutamine.

For more information about supported post-translational modifications, please see Peptide Modifications.

UniProt Integration¶

When GlycResoft digests a protein, it parses the definition line looking for accession code, and tries to guess if your definition contains a UniProt accession. If it thinks so, it will query UniProt’s web service or a local annotation file if provided for additional information. Currently, only additional cleavage sites are used to predict non-specific cleavage events at annotated locations. This is especially useful when a signal peptide cleavage site is relatively close to a glycosylation site.

If you do not wish for annotations to be applied, pass - for --annotation-path.

glycresoft documentation