Xenopus reference

From Marcotte Lab
Revision as of 12:18, 9 January 2013 by Marcotte (talk | contribs) (→‎Version 3. With Gilchrist EST set)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

Version 3. With Gilchrist EST set

EST data are courtesy of Mike Gilchrist at National Institute for Medical Research, UK. You can directly access his assembled EST data (both X. laevis and X. tropicalis) at http://genomics.nimr.mrc.ac.uk/online/xenopus/ .

Data processing method

  1. Download Mike Gilchrist's assembled EST collection (http://genomics.nimr.mrc.ac.uk/online/xenopus/Xl4-ASMs-ALTs-plus-TOP-other-16sep11-SOLID.fasta) (44,585 sequences; download on Sep., 16, 2011). Because some assembled ESTs are fragmented (more than one sequence for predicted transcript), total number of predicted transcripts is 32,859.
  2. Combine it to 'v2_all' sequences after renaming header (just add 'mgEST_Xl4_' prefix), to make 'v3_all' sequences (69,684 sequences).
  3. Sort v3_all sequences with length.
  4. Use non-redundant v2_nr080 sequences as database, and run 'usearch' with DB search + clustering mode. As a result, all 'v3_all' sequences would be either (1) clustered with v1_nr080 sequence if their %id is greater than 80%, or (2) formed as an independent cluster if there is no sequences available in v1_uc080 with %id>80%.
  5. Combining all sequences in v2_uc080 (although some of them do not have hits in search, they are all included in v3 sequences) and seed sequences of each clusters (longest sequences in each clusters), make v3 sequences ('v3_nr080' for %id>0.80 and 'v3_nr090' for %id>0.90).

Files

  • v3_nr080 cDNA: XENLA_cDNA.v3_nr080.fasta (30,692 non-redundant cDNA sequences)
    • The format of headers:'gene_name | GenBank accession | cluster size | average %similarity of clusters' (if a sequence in database cannot form a cluster, it has extra '|NotAdded' entity at the end of header line.)

Supplementary files

Version 2. With non-RefSeq cDNA

Data processing method

  1. Cluster 'v1_all' cDNA sequences with usearch(version 4.2.66) with %identity>80%, to remove redundancy (See TXGP_ens63_reference for more information about this %id cutoff). As a result, we got 8,164 non-redundant cDNA sequences (called 'v1_nr080'). Although we don't use it for further analysis, we also made non-redundant cDNA sequences with %identity>90%, more stringent criteria ('v1_nr090').
  2. From xlaevisMRNA.fasta file to be used in version 1, collect all sequences that were not included in v1_all, because (1)they have no designated gene name, or (2) they are not annotated as RefSeq (GenBank submitted sequences). It contains 16,220 sequences (called 'v2_new').
  3. Combine it to 'v1_all' sequences, to make 'v2_all' sequences (25,099 sequences).
  4. Sort v2_all sequences with length.
  5. Use non-redundant v1_nr080 sequences as database, and run 'usearch' with DB search + clustering mode. As a result, all 'v2_all' sequences would be either (1) clustered with v1_nr080 sequence if their %id is greater than 80%, or (2) formed as an independent cluster if there is no sequences available in v1_uc080 with %id>80%.
  6. Combining all sequences in v1_uc080 (although some of them do not have hits in search, they are all included in v2 sequences) and seed sequences of each clusters (longest sequences in each clusters), make v2 sequences ('v2_nr080' for %id>0.80 and 'v2_nr090' for %id>0.90).

Files

  • v2_nr080 cDNA: XENLA_cDNA.v2_nr080.fasta (9,941 non-redundant cDNA sequences; previously called 'XENLA_cDNA_ref.v2.fasta' with minor changes.)
    • The format of headers:'gene_name | GenBank accession | cluster size | average %similarity of clusters' (if a sequence in database cannot form a cluster, it has extra '|NotAdded' entity at the end of header line.)

Supplementary files

Version 1. RefSeq

Data processing method

  1. Download 'NcbiMrnaXenbaseGene_laevis.txt' and 'xlaevisMRNA.fasta' from XenBase (downloaded on May, 01, 2011).
  2. Read gene name for each NCBI id from 'Ncbi...' file. Filter out genes with 'unnamed' in gene name field.
  3. Read all sequences from '.fasta' file. Convert all sequence character to upper case.
  4. If a sequence has '>gi|<gi number>|ref|<genbank accession>' header (means it is RefSeq entity), report it.

Files

  • cDNA: XENLA_cDNA.v1_all.fasta (8,879 sequences; previously called 'XENLA_cDNA_ref.v1.fasta')
  • proteins: XENLA_prot.v1_all.fasta (8,878 sequences; 'taf5' is not annotated as RefSeq in protein, although its corresponding mRNA sequence is annotated as RefSeq.)