| Back to PME-5 main page |
This page tries to explain any doubt that you have related to the experiment, the results template, the MIAPE generation…etc. |
| |
About the results template |
| |
|
|
|
|
|
|
|
|
|
| |
|
Questions: |
|
|
|
|
|
| |
|
|
|
|
|
|
|
|
|
| |
|
|
|
|
|
|
|
|
|
| |
|
|
|
|
|
|
|
|
|
| |
|
|
|
|
|
|
|
|
|
| |
|
|
|
|
|
|
|
|
|
| |
|
|
|
|
|
|
|
|
|
| |
|
|
|
|
|
|
|
|
|
| |
|
|
|
|
|
|
|
| |
|
Responses: |
|
|
|
|
|
| |
|
What is the meaning of "unique peptide"? |
| |
|
|
- Ej: HPPEDIMLVLEEK and HPPEDIM*LVLEEK are the same, so it should be counted once.
- Ej: HPPEDIMLVLEEK from a 2+ charged spectra and HPPEDIMLVLEEK from a 3+ charged spectra are the same, so it should be counted once.
- Ej: AMRTK and AMRTKPLIMR are NOT the same.
|
| |
|
|
|
| |
|
What is the meaning of "number of MS/MS spectra assigned"? |
| |
|
|
- This is the number of spectra that matches to a certain peptide sequence. If several spectra matches to the same peptide sequence, all of them should be counted.
|
| |
|
|
|
| |
|
What is the meaning of "Number of E Coli single hit- proteins id"? |
| |
|
|
- This is the number of proteins identified by only one peptide. It is a kind of measure for the quality of the data used to the identifications.
|
| |
|
|
|
|
|
| |
|
In my iTRAQ or ICPL experiment I have sample A and sample B in the same run. Why do I have to fill one column for each sample in the "results summary" tab? |
| |
|
|
- It is true. In one single run the spectra that you will obtain will belong to both sample A and sample B. So it is redundant to state number of proteins in sample A in one column and the number of proteins in sample B in other column.
|
| |
|
|
- Solution: we have changed this results template:
|
| |
|
|
 |
| |
|
|
|
| |
|
|
 |
| |
|
|
|
| |
|
|
This modified template can be downloaded here: http://estrellapolar.cnb.csic.es/proteored/Biblioteca/Ficheros/933976242_PME5_Template5_2.xls |
| |
|
|
|
| |
|
Is there any other results template for DIGE experiments? |
| |
|
|
|
| |
|
|
|
| |
|
How can I calculate the False Discovery Rate from my results? |
| |
|
|
- You should perform the search using a decoy database, knowing what prefix have the decoy proteins. In our case, "rnd".
- You should calculate the FDR as: Protein Hits in Decoy database / Total Protein Hits.
- You have to sort your results by a certain significance value (a score, a p-value, ...):
|
| |
|
|
|
Rank |
AC |
Score |
decoy? |
FDR(%) |
1 |
P12312 |
100 |
no |
0 % |
2 |
P2343 |
98 |
no |
0 % |
3 |
P23423 |
96 |
no |
0 % |
4 |
P23423 |
96 |
no |
0 % |
5 |
P23233 |
93 |
no |
0 % |
6 |
P276423 |
92 |
no |
0 % |
7 |
P10990 |
79 |
no |
0 % |
8 |
P888212 |
75 |
no |
0 % |
9 |
P63626 |
71 |
no |
0 % |
10 |
P827721 |
70 |
no |
0 % |
11 |
rnd_P8213 |
51 |
yes |
1/11 = 9 % |
12 |
P12312 |
50 |
no |
1/12 = 8,3 % |
13 |
P32331656 |
31 |
yes |
1/13 = 7,7 % |
14 |
P$58487 |
30 |
no |
1/14 = 7,1 % |
15 |
P8435 |
27 |
no |
1/15 = 6,6 % |
16 |
rnd_P899934 |
22 |
yes |
2/16 = 12,5 % |
17 |
rnd_P8485834 |
14 |
yes |
3/17 = 17,6 % |
18 |
P46656 |
12 |
no |
3/18 = 16,6 % |
19 |
P040045 |
11 |
no |
3/19 = 15,7 % |
|
| |
|
|
- Then, since you know what proteins come from the decoy database, you only have to count them and calculate the FDR.
- Then, you have to choose all proteins with an associated FDR value below from a certain given value. For example, in this example, if we choose a FDR < 10%, we will take 15 proteins. If we choose a FDR < 18%, we will take 19 proteins.
|
| |
|
|
|
| |
|
|
|
| |
|
|
|
| |
|
|
|
| |
|
|
If you have any other doubt, please contact to smartinez[]proteored.org in order to include it in this page |
| |
|
|
|
|
|
|
|
|
|
| Back to PME-5 main page |
