ProtParts

ProtParts clusters and partitions biological protein sequences for machine learning.

Installation

pip install -r requirments.txt

Usage

$ python protparts.py -h
usage: protparts.py [-h] -i INPUT_FILE [-c THRESHOLD_C] [--exps EXP_S]
                    [--expe EXP_E] [-r THRESHOLD_R] [-p NUM_PARTITIONS]
                    [-f {JSON,TXT,CSV,FASTA}] -o OUTPUT_DIR [--prune]
                    [--makeblastdb MAKEBLASTDB_EXEC] [--blastp BLASTP_EXEC]
                    [--tmpdir TMP_DIR]

Protein clustering and partitioning

optional arguments:
  -h, --help            show this help message and exit
  -i INPUT_FILE         Input fasta file
  -c THRESHOLD_C        Threshold for clustering (use comma , to separate multiple thresholds)
  --exps EXP_S          Starting exponent for threshold
  --expe EXP_E          Ending exponent for threshold
  -r THRESHOLD_R        Threshold for sequence redundancy reduction.
                        None: skip redundancy reduction
                        (Default: None)
  -p NUM_PARTITIONS     Number of partitions. 0: skip partitioning
  -f {JSON,TXT,CSV,FASTA}
                        Output format
                        (Default: JSON)
  -o OUTPUT_DIR         Output directory
  --prune               Pruning clusters to improve clustering performance
  --makeblastdb MAKEBLASTDB_EXEC
                        Path to makeblastdb executable
                        (Default: config.MAKEBLASTDB_EXEC)
  --blastp BLASTP_EXEC  Path to blastp executable
                        (Default: config.BLASTP_EXEC)

Clustering with a threshold

python protparts.py -i example.fa -c 1e-9 -o results/

Clustering with multiple thresholds

python protparts.py -i example.fa -c 1e-8,1e-9,1e-10 -o results/

Clustering with a range of thresholds in exponent, for exmaple, using threshold between 1e-5 and 1e-10.

python protparts.py -i example.fa --exps 5 --expe 10 -o results/

Performance sequence redundancy reduction with a threshold before clustering

python protparts.py -i example.fa -c 1e-9 -r 1e-100 -o results/

Clustering and partitioning

python protparts.py -i example.fa -c 1e-9 -p 5 -o results/

Clustering based on the number of partitions This option will use the highest E-value threshold (lowest sequence homology) which can fit into the partition capacity for the given number of partitions

python protparts.py -i example.fa -p 5 -o results/

Clustering with a threshold and prune the result clusters to improve clustering performance

python protparts.py -i example.fa -c 1e-9 --prune -o results/

Output with specific output format

python protparts.py -i example.fa -c 1e-9 -f FASTA -o results/

Speicify BLAST programs and temporary directory

python protparts.py -i example.fa -c 1e-9 -o results/ --makeblastdb blast_program_dir/makeblastdb --blastp  blast_program_dir/blastp --tmpdir your_dir/tmp

Results

ProtParts will create a report of clustering result in html format under the result directory, which contains parameters for clustering and partitioning, stastical description of clusters, and graphical analysis of clusters.

Output format

JSON

The JSON has python dictionary-like format. The sequence ID can be accessed by partition or cluster index.

{
    "Cluster_0": [
        "A0002",
        "A0003",
        "A0004",
        ...

or

{
    "Partition_0": {
        "Cluster_4": [
            "A0010",
            "A0118",
            "A0119",
            ...

TXT

The TXT file contains basic information of the clustering results, starting with #. The ClustID 0 indicates the numbering of clusters, which is followed by sequence ID. If the number of partitions is provides, extra information PartID 0 will be appended after the ClustID, showing the partition numbering.

# Clustering method: graph
# Threshold: 1e-09
# Number of clusters: 2030
ClustID 0 A0002
ClustID 0 A0003
ClustID 0 A0004
...

or

# Clustering method: graph
# Threshold: 1e-09
# Number of partitions: 5
ClustID 4 PartID 0 A0010
ClustID 4 PartID 0 A0118
ClustID 4 PartID 0 A0119
...

CSV

The CSV file consists of SequenceID, ClusterID or optional PartitionID.

SequenceID,ClusterID
A0002,0
A0003,0
A0004,0
...

or

SequenceID,PartitionID,ClusterID
A0010,0,4
A0118,0,4
A0119,0,4
...

FASTA

The clustering and partitioning results are added to the description line after protein IDs in FASTA file.

>A0002 Cluster_0
MAQLTLLLLSLFLTLISLPPPGASISSCNGPCRDLNDCDGQLICIKGKCNDDPEVGTHICGGTTPSPQPGSCNPSGTLTCQGKSYPTYDCSPPVTSSTPAKLTNNDFSEGGDGGGPSECDESYHSNNERIVALSTGWYNGGSRCGKMIRITASNGKSVSAKVVDECDSRHGCDKEHAGQPPCRNNIVDGSNAVWSALGLDKNVGVVDITWSMA
>A0003 Cluster_0
MAQLTLLLLSLFFTLISLPPPGASISSCNGPCRDLNDCNGQLICIKGKCNDDPEVGTHICGGTTPSPQPGSCKPSGTLTCQGKSYPTYDCSPPVTSSTPAKLTNNDFSEGGDGGGPSECDESYHSNNERIVALSTGWYNGGSRCGKMIRITASNGKSVSAKVVDECDSRHGCDKEHAGQPPCRNNIVDGSNAVWSALGLDKNVGVVDITWSMA
>A0004 Cluster_0
MAQLTLLLLSLFLTLISLPPPGASISSCNGPCRDLNDCDGQLICIKGKCNDDPEVGTHICGGTTPSPQPGGCNPSGTLTCQGKSYPTYDCSPPVTSSTPAKLTNNDFSEGGDGGGPSECDESYHSNNERIVALSTGWYNGGSRCGKMIRITASNGKSVSAKVVDECDSRHGCDKEHAGQPPCRNNIVDGSNAVWSALGLDKNVGVVDITWSMA
...

or

>A0010 Cluster_4 Partition_0
MARPSFLSLVSLSLLVLSHSSAANRQPSKYQQQQKGECQIQRLNAQEPQQRIQAEAGVTEFWDWTDDQFQCAGVAACRNMIQPRGLLLPSYTNAPTLIYILKGRGITGVMIPGCPETYQSSQQSREGDVSHRQFRDQHQKIRRFQQGDVIALPAGVAHWCYNDGDSDLVTVSVEDTGNRQNQLDNNPRRFFLAGNPQQQQKEMYAKRPQQQHSGNVFRGFDTEVLAETFGVDMEMARRLQGKDDYRGHIIQVERELKIVRPPRTREEQEQQERGERDNGMEETICTARLVENIDNPSRADIFNPRAGRLTSVNSFNLPILNYLRLSAEKGVLYRNALMPPHWKLNAHCVLYATRGEAQMQIVDQRGEAVFNDRIREGQLVVVPQNFVVMKQAGNQGFEWVAIKTNENAMFNTLAGRTSALRAMPVDVLANAYQISQSEARRLKMGREEAVLFEPRSEGRDVD
>A0118 Cluster_4 Partition_0
PPTKFSFSLFLVSVLVLCLGFALAKIDPELKQCKHQCKVQRQYDEQQKEQCVKECEKYYKEKKGREREHEEEEEEWGTGGVDEPSTHEPAEKHLSQCMRQCERQEGGQQKQLCRFRCQERYKKERGQHNYKREDDEDEDEDEAEEEDENPYVFEDEDFTTKVKTEQGKVVLLPKFTQKSKLLHALEKYRLAVLVANPQAFVVPSHMDADSIFFVSWGRGTITKILENKRESINVRQGDIVSISSGTPFYIANNDENEKLYLVQFLRPVNLPGHFEVFHGPGGENPESFYRAFSWEILEAALKTSKDTLEKLFEKQDQGTIMKASKEQVRAMSRRGEGPKIWPFTEESTGSFKLFKKDPSQSNKYGQLFEAERIDYPPLEKLDMVVSYANITKGGMSVPFYNSRATKIAIVVSGEGCVEIACPHLSSSKSSHPSYKKLRARIRKDTVFIVPAGHPFATVASGNENLEIVCFEVNAEGNIRYTLAGKKNIIKVMEKEAKELAFKMEGEEVDKVFGKQDEEFFFQGPEWRKEKEGRADE
>A0119 Cluster_4 Partition_0
MGPPTKFSFSLFLVSVLVLCLGFALAKIDPELKQCKHQCKVQRQYDEQQKEQCVKECEKYYKEKKGREREHEEEEEEWGTGGVDEPSTHEPAEKHLSQCMRQCERQEGGQQKQLCRFRCQERYKKERGQHNYKREDDEDEDEDEAEEEDENPYVFEDEDFTTKVKTEQGKVVLLPKFTQKSKLLHALEKYRLAVLVANPQAFVVPSHMDADSIFFVSWGRGTITKILENKRESINVRQGDIVSISSGTPFYIANNDENEKLYLVQFLRPVNLPGHFEVFHGPGGENPESFYRAFSWEILEAALKTSKDTLEKLFEKQDQGTIMKASKEQIRAMSRRGEGPKIWPFTEESTGSFKLFKKDPSQSNKYGQLFEAERIDYPPLEKLDMVVSYANITKGGMSVPFYNSRATKIAIVVSGEGCVEIACPHLSSSKSSHPSYKKLRARIRKDTVFIVPAGHPFATVASGNENLEIVCFEVNAEGNIRYTLAGKKNIIKVMEKEAKELAFKMEGEEVDKVFGKQDEEFFFQGPEWRKEKEGRADE
...