全3件 ( 1 〜 3 )　 前 | 次 　1/ 1. /1

(無題) 削除/引用 No.5245-3 - 2016/07/16 (土) 18:20:59 - asan >For instance, if I submit 10 ug of total RNA from wild type and 2 ug of RNA from knockout, how are they normalized? In this case, I think you might wanted to use 2 ug for both samples to make sure that the initial condition is as much the same as possible. However, if your samples are very small subpopulation from mouse, you might need to do regardless of the absolute amount. As you would know, RNA seq data quality often depends on the quality (purity) of the RNA and also how much read you read. So, if you are comparing the low-read transcript, you might not get the real values. You might want to check the filter applied during the data processing. If you are handling raw data , you should check the quality of each sample first. You could ask your collaborators (if you are not taking your data youseif) if the quality of your samples were good enough. Normalization is sometimes packaged together with the conversion into RPFKM and this process might worsen your results because your samples have large difference in the initial RNA amounts. For such large data, we usually normalized with the median values based on the fact that the most of the transcripts patterns (including house keeping genes) were not altered. You might want to check how all the data are processed to get the relative differences. If you have replicate data, you can easily cross-check the result for correlation, the rank numbers of the transcripts and so on. If these data are not fairly correlated, it is possible that there are some technical errors during the data processing.

(無題) 削除/引用 No.5245-2 - 2016/07/16 (土) 12:31:48 - おお How are they normalized?... How did you normalize? Mostly, global normalization is used for this kind of analyses. Therefore, the simplest formula for normalization would be: reads of your interested gene/ the sequence depth Even if they say it is more reliable, they have not examined whether reads reflect the number of each gene transcript. So, you should validate the results if you need to describe increase or decrease in particular genes.

RNA sequence 削除/引用 No.5245-1 - 2016/07/16 (土) 05:00:32 - Ok Hi, I am wondering if someone here may give me comments. I am working to validate RNA seq data by qRT-PCR. But unfortunately, most of scored genes were not significantly changed based on qPCR. I encountered this weird results twice so far. (I mean I repeated RNA seq twice) I compare two sets of samples (2 from wild type vs 2 from knockout). Gene knocked-out was of course scored as a significantly reduced gene in knockout compared with wild type. I read a book to get general idea on RNA seq. Based on the book, RNA seq counts RNA reads and it is more reliable than RNA expression microarray. Also it said that due to its high reliability, it is not always needed to validate the results by qRT-PCR. I am thinking why i could not validate RNA seq results by qRT-PCR and I got a thought for the reasons. Because of the principle of RNA seq (counting absolute number of transcripts), is it the possible reason that significant differences in RNA amounts which I submitted between wild type and knockout affects the results? For instance, if I submit 10 ug of total RNA from wild type and 2 ug of RNA from knockout, how are they normalized? Any comments are welcome and appreciated if you can help. Best,

全3件 ( 1 〜 3 )　 前 | 次 　1/ 1. /1

パスワードを入力してチェックした記事を

チェックした記事を

このトピックにメッセージを投稿する 名前  メール  　アドレス非公開    タイトル  本文       設定  クッキーを保存（次回の入力の手間を省けます） 上に上げない（トピックの一覧で一番上に移動させません） 解決（問題が解決した際にチェックしてください） 暗証  半角英数字8-12文字の暗証番号を入れると、あとで削除、修正ができます。 送信

---

〔使い方〕 「アドレス非公開」をチェックすれば、自分のメールアドレスを公開しないで他の方からメールを受け取れます。 問題が解決した際には、解決ボタンをチェックして解決した旨のコメントをつけてください。これは、初めにトピックを作った人と管理人のみが可能です。 半角カタカナ、機種依存文字(全角ローマ数字、○の中の数字等)は文字化けの原因となりますので使わないでください。

---