Index of /wrpearson/fasta/PUBS/gonzalez09a/refprotdom1.1
29-March-2010
RefProtDom
The files in this directory contain a set of highly divergent families
that were originally selected from Pfam (v.21) and then manually
curated to supplement the homologies identified in Pfam (M. W. Gonzalez
and W. R. Pearson. RefProtDom: A Protein Database with Improved Domain
Boundaries and Homology Relationships. Manuscript in Preparation).
Six types of files are provided:
1. Reference library sequence files
2. Annotation files (list homologous domains in the reference libraries)
3. Supplementary Annotation Files
4. A tar-gzip file with sets of query sequences
5. A tar-gzip file with the trees and multiple sequence alignments for the superfamilies
6. A file of the most frequently-asked questions (FAQ.txt)
================ 1. Reference libraries ================
library_all_domains.fa.gz - Full-length Uniprot proteins
containing homologs to the query domains.
library_all_domains_rdm.fa.gz - Random-shuffles of each of
the full-length Uniprot proteins in library_all_domains.fa.gz.
library_long_domains.fa.gz - A subset of the library_all_domains.fa.gz
library from which proteins with homologous domains less than 75% of
the Pfam model length are excluded.
library_long_domains_rdm.fa.gz - Random-shuffles of each of
the full-length Uniprot proteins in library_long_domains.fa.gz
================ 2. Annotation files ==================
family_members.annot.gz - lists the domains in each sequence in the
library_*_domains.fa.gz files.
Format:
>[source]|[accession]|[sequence_name]
[superfamily][domain_start][domain_end][e-value][mode][long_domain][non_redundant]
Examples:
>up|P53627|ABFA_STRLI
PF06964 293 494 1.3e-104 pf21ls 1 1
>pfam21|P19801|ABP1_HUMAN
PF01179 296 715 2.3e-35 ua_pws 1 1
CL47 39 125 3.3e-29 pf21ls 1 1
CL47 141 241 1.3e-24 pf21ls 1 1
source "up" if the sequence matches the current (12/2009) Uniprot version of the sequence
"pfam21" if the exact sequence is no longer in uniprot, the pfam v.21 sequence is used
superfamily The pfam accession name (PF##### when the family was the sole representative of its superfamily/clan) or clan number (CL[clan_id] when the superfamily has several families that have been coalesced into one homologous group)
domain_start Sequence coordinate where domain starts
domain_end Sequence coordinate where domain ends
e-value The score of the comparison between the sequence fragment from domain_start to domain_end against the HMM model of the given family or the e-value generated by the supplemental annotation methods described in (Gonzalez, M.W. and W.R. Pearson, 2010b)
mode The type of pfam HMM model used to identify the given domain or the supplemental annotation method described in (Gonzalez, M.W. and W.R. Pearson, 2010b)
"pf21ls" mode domains match the entire footprint on the pfam domain model
"pf21fs" domains are usually fragments that only partially match the pfam domain model
"ext" are domains that were previously annotated as partial homologies whose coordinates we extended
"ua_str" mode refers to previously missed homologs found by Gonzalez and Pearson using structural evidence
"ua_rec" mode refers to previously missed homologs found by Gonzalez and Pearson using reciprocal PSI-BLAST searches
"ua_pws" mode refers to previously missed homologs found by Gonzalez and Pearson using pair-wise searches.
long_domain "0" if the sequence contains domains whose lengths are <75% of the Pfam model length. long_domain=0 sequences are only found in library_all_domains.fa.gz
"1" if the sequence only contains domains whose lengths are >=75% of the Pfam model length. long_domain=1 sequences are found in library_long_domains.fa.gz and in library_all_domains.fa.gz
non_redundant Useful to calculate family size
"0" flags a redundant domain that overlaps with another with longer sequence homology annotation
"1" flags the non-redundant domain with the longer sequence homology annotation
======================= 3. Supplementary Annotation files ==================================
pfam_to_clan.txt - Lists the pfam family to clan superfamily correspondence.
Note: The annotations on this database are at the superfamily level, which we recommend for homology evaluation. See the FAQ.txt and (Gonzalez and Pearson, NAR, 2010) for more details of why coalescing superfamilies is the preferred choice when evaluating homology.
refprotdom_domain_bound_ext.txt - Lists the domains that in pfam v.21 were annotated
as partial homologies whose coordinates we extended. Current uniprot accessions
and sequence ids are provided, as well as the corresponding pfam v.24 coordinates
Seq_id Sequence identifier
Status "available" if sequence is available in current Uniprot and in pfam24
"dropped_by_p24" if sequence is available in current Uniprot but not in pfam24
"demerged" if sequence is available in current Uniprot but has been broken into several shorter sequences
Pfam Pfam name
R_start/R_end RefProtDom domain boundaries
SS_start/SS_end SSEARCH domain boundaries
GL_start/GL_end GLSEARCH domain boundaries
v21_start/v21_end Pfam v.21 domain boundaries
v24_start/v24_end Pfam v.24 domain boundaries
Fixed_by_24 "1" if Pfam v.24 boundaries are within 10% of RefProtDom boundaries
Fixed_annot_cov The fraction of domain overlap between Pfam v.24 and RefProtDom boundaries
refprotdom_unannot_homol.txt - Lists missed/unannotated homologs in Pfam v.21 that we uncovered with
reverse PSI-BLAST searches, pair-wise searches or through SCOP/CATH structural evidence.
Seq_id Sequence identifier
Evidence "ua_rec" if the homology was uncovered using reciprocal PSI-BLAST searches
"ua_pws" if the homology was uncovered using pair-wise (i.e. SSEARCH and GLSEARCH) searches
"ua_str" if the homology was uncovered with SCOP/CATH structural evidence
Pfam (Same as *ext.txt" above)
R_start/R_end (Same as *ext.txt" above)
Status (Same as *ext.txt" above)
v24_start/v24_end (Same as *ext.txt" above)
Fixed_by_v24 (Same as *ext.txt" above)
Fixed_annot_cov (Same as *ext.txt" above)
=========================== 4. Query sequences =================================
queries.tgz is a gzip-ed tar file that produces the following directories:
queries/by_difficulty/
queries/by_tree_location/
In "queries/by_difficulty/", there are two classes of query sequence files, each of which contains 50 domain sequences, in 10 different random-sequence embeddings.
hard_embedded.[1-10].fa
sampled_embedded.[1-10].fa
In addition, there is a
hard_non_embedded.fa and
sampled_non_embedded.fa file.
"hard" domains are domains that find the smallest number of related sequences after a BLASTP search. "sampled" domains were chosen at random from 640 domains selected because of their length (>200
residues in the Pfam model) and phylogenetic diversity (homologs in 2 of the 3 kingdoms of life: e.g. homologs in archaea and eukarya, or in archaea and bacteria, etc).
"queries/by_tree_location/", also contains two classes of query sequence files, each with 50 domain sequences, in 10 different random sequence embeddings. Here, the classes are "des", for queries from relatively deserted parts of the domain phylogenetic tree, and "pop", for queries from a populated region.
QUERY EMBEDDING
All queries are available as bare domains (non-embedded/ne) or flanked by artificial proteins (embedded/e#). The embeddings were created by randomly shuffling the domain as described by Gonzalez and Pearson. For each domain, 10 different embedding replicates are provided. Unless otherwise specified, the results in (Gonzalez and Pearson, NAR, 2010) are based on embedding #5.
QUERY FILE NAMES
A query is a sequence domain from a family that falls under any of the 4 types of queries described above (i.e. hard, sampled, populated/pop, deserted/des). The four types of query families are available as bare domains (non_embedded) or embedded in 10 different shuffles following the following naming format: [type]_[embedding].[e#].fa. For example: "hard_embedded.5.fa" contains 50 embedded queries from hard families and the embedding is the 5th shuffle of the domain (there are 9 alternate embeddings for the same domain).
QUERY FILE FORMAT
Query files are in FASTA format, with the description line providing information about the location of the domain, and its origin. Each query file contains 50 queries of the form:
>qPF00589_e5 e_d_start:96 e_d_end:286 from:up|Q1YWW7|Q1YWW7_PHOPR(194-384); pfam:PF00589; model_len:205; all_homol:1445; long_homol:963; descr:Phage integrase family
KTKKSAKQSDL.... [sequence] ....
The format of the description line is:
>[query_accession] e_d_start:# e_d_end:# from:[sequence_id]([domain_start]-[domain_end]); pfam:[pfam_superfamily]; model_len:[#]; all_homol:[#]; long_homol:[#]; descr:[description]
query_accesion Accession number for each query in the format:
q[pfam_superfamily]_[ne|e#]. For example: qPF00589_e5 is a domain from PF00589 that has been embedded in the 5th shuffle replicate.
e_d_start/ The boundaries of the Pfam domain in the query sequence
e_d_end
from The original pfamseq_id (Uniprot id) and coordinates (start-end) of the query domain.
pfam The pfam accession name (PF#####) or clan number (CL###). No clan accession is given when a superfamily contains a single Pfam family.
model_len The length of the pfam domain model
all_homol Number of homologs that this family has in the "library_all_domains.fa" library
long_homol Number of homologs that this family has in "library_long_domains.fa"
descr Description of the Pfam domain
======================= 5. Trees and MSAs ==================================
trees.tgz is a gzip-ed tar file that produces the following directories:
trees/all_domains_in_family/
trees/long_domains_in_family/
"trees/all_domains_in_family/" contains trees of all domain members of each superfamily
"trees/long_domains_in_family/" contains trees of the long-domain members of each superfamily
All *.tree files in the "trees/" folder are newick formatted, neighbor-joining trees of a set of members for each superfamily (i.e. trees that feature all domain members in a superfamily or trees of only the long-domain members).
The .*afa files contain the multiple sequence alignments used to generate the trees
All superfamily trees were generated using Quicktree (v. 1.1), and the multiple sequence alignments to generate them were created using the HMMMER (v. 2.3.2) package.
================================================================
6. Using the files -- evaluating search alignment accuracy
To determine whether the alignments are True positives (TPs) or False Positives (FPs) all you need to know is the library sequence's id (e.g. up|Q1YWW7|Q1YWW7_PHOPR) and the pfam superfamily to which the query belongs (e.g. qPF00589_e5's superfamily is PF00589).
Find the library sequence in the "family_members.annot.gz" file (alternatively this information may be stored in mySQL tables) and compare the domain boundaries there to the alignment coordinates of the similarity-searching algorithm you are testing. For instance, let's assume you're testing the PF00589 superfamily (using qPF00589e5:a query from hard_embedded.5.fa) and suppose your algorithm finds a putative homolog on the "XERC_BACSU" sequence from residues 10-80. Looking at the "family_members.annot" file you would classify this alignment as a false positive (FP) because the alignment maps 100% to the unrelated PF02899 domain. You may decide to use a specific overlap percentage to classify the alignments. In (Gonzalez MW, Pearson WR: Homologous Over-extension: A Challenge for Iterative Similarity Searches. Nucleic Acids Res 2010, [Jan. 10 Epub ahead of print] doi:10.1093/nar/gkp1219:1-13), we require 50% alignment overlap to the homologous region (i.e. at least 50% of the alignment must be between 114-291) for the alignment to be counted as a true positive (TP).
>up|P39776|XERC_BACSU
PF02899 8 91 1.6e-26 ls 1 0
PF00589 114 291 1.6e-65 ls 1 1
================================================================
For more information, contact Bill Pearson (pearson@virginia.edu) or Mileidy Gonzalez (mileidygwgonzalez@gmail.com)