November 21, 2021 [reading notes] - bioinformatics and functional genomics (Chapter 5 advanced database search)

• 
• 5.3 Look for distant related proteins ： Location specific iteration BLAST（PSI-BLAST） and DELTA-BLAST
  • PAM250 The matrix provides a better scoring system for detecting distant related proteins , The scoring matrix can be changed to detect distant proteins , But there are still limitations ：
    • BLASTP A matching protein was detected , But whether it is homologous is not clear .
  • PSI-BLAST
    • Search the database more deeply , To find a matching protein that is distantly related to the protein you are interested in .
    • Comparable 5 A step ：
      • Use a scoring matrix to search the query sequence in a target database （ As usual BLASTP）
      • Construct a multiple sequence alignment from the results of the initial search based on component statistics , Based on the multiple sequence alignment , Build specialized 、 Personalized search matrix （ Or spectrum profile）
      • The resulting position specific scoring matrix （PSSM） Will be used for another search
      • PSI-BLAST Evaluate the statistical significance of database matching .
      • The above search is iterated , Generally, it iterates five times .
      • legend ： The resulting position specific scoring matrix （PSSM）

        • 

      • legend ：3 Comparison results after iterations

        • 

    • PSI-BLAST Error of ： pollution problem
      • The main source of errors is the false amplification of unrelated sequences , There are three measures to stop pollution
        • 1. A filtering algorithm is used to remove the amino acid regions composed of preferences
        • 2. Adjust the inclusion threshold from the default value to a lower value
        • 3. Visually inspect each time PSI-BLAST The result of the iteration
        • Icon ： All globulins and PSI-BLAST How to improve search sensitivity

          • 

    • Reverse position specificity BLAST（RPS-BLAST）
      • RPS-BLAST You can combine a query protein sequence with a predefined PSSM Database comparison , The conserved protein domains in the query sequence can be identified .
        • Icon ：RPS-BLAST Inquire about

          • 

  • DELTA-BLAST
    • NCBI The most sensitive and accurate protein search tool on the
    • advantage
      • Based on high quality manual audit CDD database , Can produce ratio PSI-BLAST Bigger 、 More complete PSSM.
      • Than BLASTP and PSI-BLAST More sensitive , Including more sensitive search for distant related proteins
      • Fast
      • Than BLASTP The quality of sequence alignment is better .
  • PSI-BLAST and DELTA-BLAST Performance evaluation of
    • DELTA-BLAST Sensitivity due to PSI-BLAST、BLASTP And other procedures ;
    • Under a given number of false positive conditions ,DELTA-BLAST It can be found that three times BLASTP Found homologous proteins
  • Pattern matching initiated BLAST（PHI-BLAST, pattern recognition BLAST）
    • A protein may contain an amino acid residue pattern or “ Characteristic signal ” Amino acid residues , This can help us judge whether the protein belongs to a certain family .PHI-BLAST We can find the matching results that contain both matching patterns and query sequences ;DELTA-BLAST High sensitivity , However, it does not output information about the schema selected by the user .
    • Definition of characteristic signals and modes
      • Customize , And introduce a certain degree of fuzziness , such as NDFX（5）GXW[YF]:
        • X（5） Represents that these five positions can be any kind of amino acid residue ;
        • [YF] It indicates that the amino acid residue at the last position must be one of the lysine or phenylalanine ;
        • The selected mode should not appear too often , The algorithm only allows the occurrence frequency less than 1/5000.
    • Icon ： Select a mode to PHI-BLAST Search for

      • 

    • PHI-BLAST The algorithm is based on spanning the input pattern and its upstream and downstream A1 Areas and A2 Double sequence alignment from region A0 Analyze , The result of sequence alignment is scored by the extension with vacancy .
•   Welcome to join the group , Or add VX：bbplayer2021, Invite in
•