Á¤º¸Ã³¸®ÇÐȸ ³í¹®Áö D
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
SVM ¸ðµ¨À» ÀÌ¿ëÇÑ 3Â÷¿ø ÆÐÄ¡ ±â¹Ý ´Ü¹éÁú »óÈ£ÀÛ¿ë »çÀÌÆ® ¿¹Ãø±â¹ý |
¿µ¹®Á¦¸ñ(English Title) |
Prediction of Protein-Protein Interaction Sites Based on 3D Surface Patches Using SVM |
ÀúÀÚ(Author) |
¹Ú¼ºÈñ
Bjorn Hansen
Sung Hee Park
Bjorn Hansen
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 19-D NO. 01 PP. 0021 ~ 0028 (2012. 02) |
Çѱ۳»¿ë (Korean Abstract) |
¸ð³ë¸Ó ´Ü¹éÁúÀÇ »óÈ£ÀÛ¿ë »çÀÌÆ® ¿¹ÃøÀº ±â´ÉÀ» ¾ËÁö ¸øÇÏ´Â ´Ü¹éÁú¿¡ ´ëÇؼ ÀÌ°Í°ú »óÈ£ÀÛ¿ëÇÏ´Â ´Ü¹éÁú·ÎºÎÅÍ ±â´ÉÀ» ¿¹ÃøÇϰųª ´Ü¹éÁú µµÅ·À» À§ÇÑ °Ë»ö °ø°£ÀÇ °¨¼Ò¿¡ Áß¿äÇÑ ¿ªÇÒÀ» ÇÑ´Ù. ±×·¯³ª »óÈ£ÀÛ¿ë»çÀÌÆ® ¿¹ÃøÀº ´ëºÎºÐ ´Ü¹éÁú »óÈ£ÀÛ¿ëÀÌ ¼¼Æ÷ ³»¿¡¼ ¼ø°£Àû ¹ÝÀÀ¿¡ ÀϾ´Â ¾àÇÑ »óÈ£ÀÛ¿ëÀ¸·Î ½ÇÇè¿¡ ÀÇÇÑ 3Â÷¿ø °áÁ¤ ±¸Á¶ ½Äº°ÀÇ ¾î·Á¿òÀÌ µû¸£¸ç ÀÌ·Î ÀÎÇØ 3Â÷¿øÀÇ º¹ÇÕü µ¥ÀÌÅÍ°¡ Á¦ÇÑÀûÀ¸·Î ¾ç»êµÈ´Ù.
ÀÌ ³í¹®¿¡¼´Â ¸ð³ë¸Ó ´Ü¹éÁúÀÇ 3Â÷¿ø ÆÐÄ¡ °è»êÀ» ÅëÇÏ¿© ±¸Á¶°¡ ¾Ë·ÁÁø º¹ÇÕüÀÇ »óÈ£ÀÛ¿ë»çÀÌÆ®¿Í ºñ»óÈ£ÀÛ¿ë»çÀÌÆ®¿¡ ´ëÇÑ ÆÐÄ¡ ¼Ó¼ºÀ» ÃßÃâÇÏ°í À̸¦ ±â¹ÝÀ¸·Î Support Vector Machine (SVM) ºÐ·ù±â¹ýÀ» ÀÌ¿ëÇÑ ¿¹Ãø ¸ðµ¨ °³¹ßÀ» Á¦½ÃÇÑ´Ù. Ÿ°Ù Ŭ·¡½ºÀÇ µ¥ÀÌÅÍ ºÒ±ÕÇü ¹®Á¦ ÇØ°áÀ» À§ÇØ under-sampling ±â¹ýÀ» ÀÌ¿ëÇÑ´Ù. »ç¿ëµÈ ÆÐÄ¡¼Ó¼ºÀº 2Â÷ ±¸Á¶ ¿ä¼Ò¿Í ¾Æ¹Ì³ë»ê ±¸¼ºÀ¸·ÎºÎÅÍ ÃÑ 9°³°¡ ÃßÃâµÈ´Ù. 147°³ÀÇ ´Ü¹éÁú º¹ÇÕü¿¡ ´ëÇؼ 10 fold cross validationÀ» ÅëÇؼ ´Ù¾çÇÑ ºÐ·ù¸ðµ¨ÀÇ ¼º´É Æò°¡¸¦ ÇÏ¿´´Ù. Æò°¡ÇÑ ºÐ·ù ¸ðµ¨ Áß SVMÀº 92.7%ÀÇ ³ôÀº Á¤È®¼ºÀ» º¸ÀÌ°í À̸¦ ÀÌ¿ëÇÏ¿© ºÐ·ù ¸ðµ¨À» °³¹ßÇÏ¿´´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
Predication of protein interaction sites for monomer structures can reduce the search space for protein docking and has been regarded as very significant for predicting unknown functions of proteins from their interacting proteins whose functions are known. In the other hand, the prediction of interaction sites has been limited in crystallizing weakly interacting complexes which are transient and do not form the complexes stable enough for obtaining experimental structures by crystallization or even NMR for the most important protein-protein interactions.
This work reports the calculation of 3D surface patches of complex structures and their properties and a machine learning approach to build a predictive model for the 3D surface patches in interaction and non-interaction sites using support vector machine. To overcome classification problems for class imbalanced data, we employed an under-sampling technique. 9 properties of the patches were calculated from amino acid compositions and secondary structure elements. With 10 fold cross validation, the predictive model built from SVM achieved an accuracy of 92.7% for classification of 3D patches in interaction and non-interaction sites from 147 complexes.
|
Å°¿öµå(Keyword) |
´Ü¹éÁú »óÈ£ÀÛ¿ë
3Â÷¿ø ÆÐÄ¡
ÀÎÅÍÆäÀ̽º
¹ÙÀεù »çÀÌÆ® ¿¹Ãø
SVM
ºÐ·ù
µ¥ÀÌÅÍ ºÒ±ÕÇü
Protein Interaction
3D Patches
Interface
Prediction of Binding Sites
SVM
Classification
Imbalanced Data
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|