> Á¤º¸¼¾ÅÍ > ¼¼¹Ì³ª/Çà»ç
Á¦¸ñ | 2021³â 9¿ù 30ÀÏ »ý¹°Á¤º¸ÇÐ ¼¼¹Ì³ª | ||
---|---|---|---|
ÀÛ¼ºÀÚ | °ü¸®ÀÚ | ÀÛ¼ºÀÏ | 2021-09-23 |
¼¿ï´ëÇб³ »ý¹°Á¤º¸ÇÐ Çùµ¿°úÁ¤ ÁÖÃÖ·Î ¼¼¹Ì³ª¸¦ ¾Æ·¡¿Í °°ÀÌ ¿°íÀÚ ÇÏ¿À´Ï ¸¹Àº Âü¿© ¹Ù¶ø´Ï´Ù. ¼¼¹Ì³ª´Â ZoomÀ» ÅëÇÑ ¿Â¶óÀÎ °ÀÇ·Î ¿¹Á¤µÇ¾î ÀÖÀ¸¸ç °ÀÇ¿¡ Âü¿©ÇÏ°íÀÚ ÇϽô ºÐ²²¼´Â ¾Æ·¡ÀÇ ¸µÅ©·Î Âü¿© ºÎŹµå¸®°Ú½À´Ï´Ù. ÀϽÃ: 2021³â 9¿ù 30ÀÏ ¸ñ¿äÀÏ ¿ÀÀü 11½Ã ¿¬»ç: ³²´ö¿ì ¿¬»ç´Ô (UNIST) Zoom link:¡¡https://snu-ac-kr.zoom.us/j/84015763984 Title:¡¡Normalization, batch-effects correction and differential expression analysis of single-cell RNA-seq data Abstract Despite the recent advancement in integration of scRNA-seq data from different samples, the strategies to integrate differential expression (DE) analysis of scRNA-seq data remains underexplored. Here, we benchmarked three integrative approaches for DE analysis of scRNA-seq data: batch-effects correction followed by DE analysis, statistical model incorporating batch covariate, and meta-analysis. Batch-effects correction methods developed for scRNA-seq data rarely improved the analysis of uncorrected data, whereas other commonly used integration methods showed modest improvement. In contrast, the choice of DE analysis method had a significant impact on the performance. In particular, several parametric methods consistently outperformed the widely used ranksum test. We analyzed the scRNA-seq data for seven patients with lung adenocarcinoma. Several integration methods exhibited significant and improved ranks of known genes in lung cancer compared to the ranksum test, and also better prioritized the genes prognostic to lung cancer. |
|||