在C45Split用到了如下代码:
m_distribution = new Distribution(m_complexityIndex,
trainInstances.numClasses());//包数=分枝数
Enumeration enu = trainInstances.enumerateInstances();
while (enu.hasMoreElements()) {
instance = (Instance) enu.nextElement();
if (!instance.isMissing(m_attIndex))
m_distribution.add((int)instance.value(m_attIndex),instance);//把没有缺失值的实例的权值加入到相应(对应其划分属性的取值)的包中
private static InfoGainSplitCrit infoGainCrit = new InfoGainSplitCrit();
m_infoGain = infoGainCrit.
splitCritValue(m_distribution,m_sumOfWeights);
m_gainRatio =
gainRatioCrit.splitCritValue(m_distribution,m_sumOfWeights,
m_infoGain);
展开InfoGainSplitCrit.java的构造方法:
public final double splitCritValue(Distribution bags,double totalNoInst,
double oldEnt) {
double noU
double unknownR
noUnknown = totalNoInst-bags.total();
unknownRate = noUnknown/totalNoI
numerator = (oldEnt-newEnt(bags));
numerator = (1-unknownRate)*
if (Utils.eq(numerator,0))
return numerator/bags.total();
在EntropyBasedSplitCrit.java中，展开oldEnt()：
public final double oldEnt(Distribution bags) {
double returnValue = 0;
for (j=0;j&bags.numClasses();j++)
returnValue = returnValue+logFunc(bags.perClass(j));
return logFunc(bags.total())-returnV
划分后，newEnt展开:
public final double newEnt(Distribution bags) {
double returnValue = 0;
for (i=0;i&bags.numBags();i++){
for (j=0;j&bags.numClasses();j++)
returnValue = returnValue+logFunc(bags.perClassPerBag(i,j));
returnValue = returnValue-logFunc(bags.perBag(i));
return -returnV
logFunc()做对数运算:
protected static double log2 = Math.log(2);
public final double logFunc(double num) {
if (num & 1e-6)
return num*Math.log(num)/log2;
GainRationSplitCrit.java用来计算信息增益，构造方法：
public final double splitCritValue(Distribution bags, double totalnoInst,
double numerator){
double noU
double unknownR
denumerator = splitEnt(bags,totalnoInst);
if (Utils.eq(denumerator,0))
denumerator = denumerator/totalnoI
return numerator/
splitEnt()方法:
private final double splitEnt(Distribution bags,double totalnoInst){
double returnValue = 0;
double noU
noUnknown = totalnoInst-bags.total();
if (Utils.gr(bags.total(),0)){
for (i=0;i&bags.numBags();i++)
returnValue = returnValue-logFunc(bags.perBag(i));
returnValue = returnValue-logFunc(noUnknown);
returnValue = returnValue+logFunc(totalnoInst);
return returnV
总结：在用属性划分实例后，会出现总数为按属性值个数的几个包，计算划分属性的增益只用考虑每个包中的实例个数，不用管实例的类别。而计算划分后整体的增益就必需要考虑每个包中实例的类别。weka对缺失值的处理是采用的是缺失值实例的比重来折算属性的信息增益。
&&相关文章推荐
* 以上用户言论只代表其个人观点，不代表CSDN网站的观点或立场
访问：733次
排名：千里之外告诉我英文“split ratio”解释是什么意思及翻译？ - 百科问答
& 告诉我英文“split ratio”解释是什么意思及翻译？
告诉我英文“split ratio”解释是什么意思及翻译？
悬赏分：0 -
- 阅读次数:
告诉我英文“split ratio”解释是什么意思及翻译？
关于"告诉我英文“split ratio”解释是什么意思及翻译？"的最佳答案
告诉我英文“split ratio”解释是什么意思及翻译？：
【化】 分流比
&&&&点击左边的图片给他一个评价吧
"告诉我英文“split ratio”解释是什么意思及翻译？"由
提出。如果您知道该问题答案，欢迎注册/登录后回答。 &
对"告诉我英文“split ratio”解释是什么意思及翻译？"最佳答案的评论
评论字数在200 字以内安全检查中...
请打开浏览器的javascript，然后刷新浏览器
< 浏览器安全检查中...
还剩 5 秒&ES Cell Culture and Manipulation
来源:本站原创
浏览人数: 1714
ES Cell Culture and Manipulation
Culturing ES cells
High glucose DMEM (-pyruvate, -glutamine)
20% Heat-inactivated Fetal calf serum (can vary by cell type, be sure!!)
1X l-glutamine
1X Penicillin/streptomycin
1X Non-essential amino acids
1X ribonucleosides
1/100 volume &-ME stock (stock is 7&l &-ME in 10ml DMEM)
1:1000 dilution of LIF (add 500 &l 720LIFD cell conditioned media to 500ml media)
Thawing frozen vials of ES cells
Thaw vial rapidly at 37°C, rinse outside of vial with 70% ethanol or 1% Rocall. Wipe off.
Use a 5ml pipette to transfer cells to a l5ml centrifuge tube with 10 ml. ES cell media and spin 2 min at 2000 rpm
Aspirate and resuspend cells in fresh media (be gentle, ES cells are easily damaged after thawing).
Mic's Notes: Typically, one vial contains one confluent plate of cells, frozen down. Depending on how well the freezing works one vial is typically enough for 4-12 plates. For quick recovery, plate out less. For maximum expansion plate out more. 10 plates usually is borderline, go 1:12 ONLY if you know that the recovery will be very high.
Plating ES cells
See above for notes on splitting ES cells, frozen. Generally, ES cells are split 1:6-1:10. This depends on how "confluent" the plate is. Since ES cells grow as discrete colonies, you should NEVER get a plate that is confluent in ES cells the way fibroblasts are. You wish to split them when a colony is approximately 1-2 times the diameter of a stretched out fibroblast (STO cell). If the plate is densely covered with colonies, they may be split 1:10, or slightly greater. If there are a lot of colonies, split them 1:6-1:8. For a plate that has only a few colonies (12-20), allow them to get slightly larger than normal, and then split 1:4. To split, simply add 1ml. of the appropriate dilution of cells to a 100cm. plate containing a feeder layer, and 4 ml. of ES cell media.
Splitting ES cells
When ES cells are "confluent" and ready to split (see above) feed them with fresh media an hour or two before splitting. Aspirate media and wash cells once briefly with 1X trypsin-EDTA, aspirate. Add 0.5ml 1X trypsin-EDTA (for one well of a 6-well, 2 ml. for a 100cm plate). Then, add .5 ml. pre-warmed (37°C) trypsin to a 6-well, or 3ml. to a 100cm. and put at 37°C for 5-6 min. You wish to minimize this time, so monitor the cells carefully. When cells are loosened add 0.5ml ES cell media to the well with a 5 ml pipette. Attach a sterile yellow tip to the end of a 10 ml sterile plactic pipette and pipette cells up and down several times -- ES cells are sticky and this is necessary to achieve a single-cell suspension to prevent differentiation after plating. For a 100cm. plate, transfer the cells to a Centrifuge tube containing an equal volume of media:trypsin. Pipette back and forth at least three times using a sterile yellow tip attached to the bottom of a 10ml. pipette. Transfer cells to a l5ml centrifuge tube, add another 5 ml ES cell media and spin 2 min at 2000 rpm. For a 100cm. plate, add another trypsin volume's worth of ES cell media to the centrifuge tube, and mix. This may require the use of a 50 ml centrifuge tube. Spin 2 minutes at 2000 rpm. Aspirate and resuspend cells in ES cell media, plate onto fresh feeder cells (remember to be careful about the split ratio. ES cells will grow slowly, or not at all, if they are not dense enough. Unless the plate that you are splitting from is extremely dense, split cells no more than 1:10)
Freezing ES cells
Freeze as for STO/SNL cells, ES freezing media is regular ES media with 20% FBS and 10% DMSO. It is very important that the cells be cooled slowly, to prevent the formation of ice-crystals within the cells. This can be accomplished one of two ways. One way, is to freeze the cells at -20°C for an hour, and then transfer them to a -70°C overnight. The next day, move them to liquid nitrogen. An alternate way to freeze cells, if to use a cell-culture cooler (simply a thick styrofoam box, with a few holes poked in it). Place the cells inside the cell-culture cooler, and place at -70°C overnight. The next day, transfer the cells to liquid nitrogen.
Transformation of ES cells
Preparation of targeting DNA
Digest 100&g of targeting construct DNA with an enzyme to linearize Check 0.5&l on a gel to make sure it's completely digested. Extract DNA 2X with phenol/chlorofom. Add .15 volumes 2M NaOAc and 2 volumes EtOH to precipitate (the DNA will probably come right out of solution so freezing isn't needed). Pellet the DNA, wash with 70% EtOH and dry the pellet. Resuspend the DNA in 100&l sterile H20 or TE (you should be working in the hood here). Check 0.5 &l DNA on a gel, and make sure that the concentration is know.
Preparation of cells for transformation
For one electroporation experiment you will need approximately 6 x 107 ES cells (~3 confluent 10cm plates). Feed the cells in the morning and wait a couple hours before electroporating.
Trypsinize the cells as above and wash 2X with plain DMEM (wash by spinning and resuspending, etc.). Resuspend the cells in 10ml plain DMEM and count by hemocytometer (count cells/25 center squares x 104 x cell volume in ml = total # of cells).
Centrifuge cells and resuspend in 2.4ml DMEM, divide into 3 aliquots of 800 &l each (~2 x 107 cells/800&l. Add 30&g sterile linearized targeting DNA each to 2 tubes (1 for G418 selection, 1 for G418/gancyclovir); to the third tube add an equal volume of DNA resuspension solution (H2O or TE) for a neg. control. Transfer each cell/DNA mix to a 4mm electroporation cuvette, wait 5 min and zap at room temp as below: Bio-Rad electroporator (Bio-Rad, Inc. , CA) -- 250V, 500&F. You will need to diconnect the resistance and use the capacitance extender to achieve these settings. Monitor time constants, which should be ~7 seconds.
Let cuvettes sit 5 min at room temperature and divide each over 3 10cm feeder plates in regular ES media.
Drug Selection.
The day after electroporation there should be some floating dead cells but most stuck down on the feeders.
Media will be very yellow. Feed all plates with ES cell media containing 400&g/ml neomycin (G418, (Geneticin, Gibco #)) and 1pM gancyclovir (Syntex, Palo Alto, CA), if using HSV-TK secondary selection. You may need to adjust this concentration slightly (for neomycin), as it varies between cell lines. Change the selection media every day for the first few days until cell number decreases, then every 2 days. After 5-6 days in selection you will begin to see microscopic colonies of ES cells growing out and the background will be much reduced. Expect to see extensive cell death during the first 3-5 days of selection. Also, your cells may not look to be affected for several days after adding drug. Don’t worry, have patience. Pick the ES colonies when they are easily seen macroscopically but have not yet begun to show any signs of differentiation, this is usually after ~9-11 days in selection but some variability is not unusual. You will need large colonies, in order to have enough cells for subsequent steps of freezing and screening.
Picking ES cell clones
One or two days before picking colonies prepare 24-well plates of feeders. You can also use alternate protocols that utilize 96-well plates, small tissue culture dishes, or 6 well plates, if you prefer. To pick clones you will need: fresh trypsin-EDTA, a P-200 pipetman and sterile yellow tips (preferably barrier, or plugged tips), a sterile 96well round-bottom plate, plenty of ES cell media, and an inverted microscope in a hood. Aspirate the REF media from the 24-wells and add 2ml of ES cell media to each well. Add 40 &l trypsin-EDTA to as many wells of the 96-well plate as you plan to pick clones. Working with one 10cm plate at a time, wash the plate with 2ml trypsin and then add another 2ml. Using the P-200 pick up a small amount trypsin (~5&1) from the edge of the plate then pick up a well isolated colony in a minimal amount of trypsin. You can dislodge the colony slightly with the pipette tip. Transfer the colony to a well of the 96-well plate and number this well so you don't mix clones together. Repeat for each colony to be picked from this plate. Pick all sizes of colonies but you will have to work fast before the plate dries out. If clones become too loose and start to move around you can reduce trypsin to lml or skip the trypsin rinse. After picking the first 1/3 of the clones they must be broken up into a single-cell suspension -- this is very important because any clumps of cells will begin to differentiate, often within the first few days! Use the P-200 with yellow tips to pipette each clone many times (10-20X, try to keep number of bubbles down) and check individually under the 4X microscope to be sure no clumps are left. Transfer each disaggregated clone to a numbered 24-well feeder (the 2ml media will neutralize the trypsin). Notes: 1.) Change yellow tips for each clone at each step, you don't want to cross contaminate them 2.) The ES cells can take more pipetting and longer times in trypsin than you would expect, however work quickly and with one plate at a time so trypsin exposure doesn't become excessive 3.) I usually check all clones visually under the scope ~ every 2 days and if some are growing clumpy they may need to be split 1:1 before they are actually confluent.
Expansion of ES cell clones
Change the ES cell media of each picked clone the following day (now you only add lml per well) and every 1-2 days until the clones begin to become confluent (as evidenced by yellowing media) in ~5 days. When a clone is confluent it is trypsinized and split to one well of a fresh 24-well feeder and a gelatinized well of a 6-well plate, be sure to make the 24-well feeders ahead of time. Trypsinize by rinsing with 0.5ml trypsin-EDTA, aspirate and add another 200&l trypsin for ~10 minutes. Add 800&l ES media with a lml pipette with a yellow tip and pipette several times to break up cells. Transfer to a sterile l.5ml eppendorf tube and centrifuge ~2min at 3000rpm. Aspirate media and use a P-200 to resuspend pellet in 200,u1 ES media, transfer to feeder and gelatin wells (~140&l to feeder well and 60&1 to gelatin). Feed the 24-well feeder plate the following day and by the next day the cells should be yellowing the media and are ready to be frozen (see freezing protocol below). I don't usually refeed the gelatin plate, just scrape the cells for DNA in several days when media is yellow. Note: If you have picked many clones the above method may be impossible to follow time-wise. A way to shorten the protocol is to eliminate the centrifuge step and just add the lml trypsin/ES mix to the two plates (700&1 to feeders and 300&l to gelatin). Even this small amount of trypsin may make the clones grow a bit slower initially but if you change the media early the next day they seem to do OK and this saves a lot of time.
Freezing of ES cell clones in 24-well plates
When the 24-well feeder plates of clones begin yellowing the media they are ready to be frozen (~2 days). Make ES cell freezing media ahead of time and chill in fridge or on ice Also prechill a small styrofoam box at -80°C (the boxes NEB ships enzymes in are a perfect size, or you can hollow out the styrofoam trays from 15ml centrifuge tubes and make a sort of box out of them. This box is identical to the ES cell freezing box described above). Aspirate the media from a 24-well plate of clones and add 400&1 prechilled freezing media to each well Wrap the plate wel if freezing multiple plates place on ice until all are done. As quickly as possible move plates into prechilled box in -80° freezer. Freeze plates several hours or overnight at -80° then move box to a -135°C freezer for long-term storage (i.e. liquid nitrogen, if possible). Cells can be kept long-term at -80° but it should be in a freezer that is rarely opened since the worst possible thing is repeated freeze-thaw!
Scraping ES cells from gelatin for DNA
When cells are confluent and media is good and yellow scrape cells to make DNA. Aspirate media and carefully rinse each well 2 times with lml of incomplete PBS. Add another lml PBS and scrape cells with a rubber policeman, transfer to an eppendorf tube and freeze at -20°C until ready to extract DNA.
Gelatin treatment of Tissue-culture plates
Make a 0.1% solution of gelatin (Sigma # G-1890) in Milli-Q water and autoclave, cool. Add enough gelatin to each plate, or well of a plate, to cover the bottom and let sit at least 5 minAspirate and add STO/SNL media.
Extraction of DNA from ES cells
Rapid Preparation of DNA from ES cells in 24-well tissue culture dishes
This simple method based on a protocol described by Miller et al. (1988) involves salting out cellular proteins with a saturated NaCl solution. It does not require extraction with phenol. Sufficient DNA can be obtained by this method for screening by Southern blot analysis (3-5 &g)
Materials And Equipment
Prepared 24-well tissue culture dishes containing nearly confluent ES
cell cultures
Phosphate-buffered saline (PBS) (incomplete)
Lysis buffer (150 mM NaCl, 2 mM EDTA, 1% SDS, 20 mM Tris-HCl [pH 8.0],
&g/ml proteinase K)
microfuge tubes
Solution of saturated NaCl
and 100% ethanol at room temperature
automatic micropipettor with a disposable tip or glass rod
ml of TE buffer
Caution: Wear a mask while weighing SDS.
1. Aspirate medium from each well containing nearly confluent ES cells and rinse the cells once in PBS. Resuspend the cells in 200 yl of Iysis buffer. Incubate the dish for several hours to overnight at 55°C.
2. Transfer to 1.5-ml eppendorf tubes. After digestion with lysis buffer is complete, add 100 &l of saturated NaCl to each tube and shake vigorously (do not vortex, to avoid shearing fragile genomic DNA).
3. Centrifuge the tubes at 3000g in a microfuge for 15 minutes. Transfer the supernatant containing DNA to a fresh 1.5-ml polypropylene tube and add 2 volumes of ethanol at room temperature. Invert the tube several times until the DNA precipitates and then remove it with the disposable tip of a 200-&l pipetteman or with a glass rod. Rinse the pellet in 70% ethanol and resuspend it in 50 &l of TE. Allow the pellet to dissolve at room temperature for several hours.
Preparation of LIF Conditioned Media
The 720 LIFD cell line has been transformed with a plasmid that causes them to secrete a high level of LIF. For unknown reasons, this causes the cells to grow very slowly. They are also more easily contaminated (in my hands) than normal STO cells. Also, the cells will look sick, compared to normal STO cells. They tend to be slightly smaller, and much more vacoulated.
GIBCO a-MEM (-l-glutamine)(720LIFDs are fussy about the brand of media, they don't like Sigma)
10% Heat-inactivated calf serum
1X Penicillin/streptomycin
DO NOT add l-glutamine, the cells don't like it!
Growing 720LIFD cells
Grow cells on gelatinized 10cm tissue culture plates (see STO/SNL protocol for gelatinizing). Change media every 2 days, the cells will grow slowly, and look messed up. This is normal for this cell line. Split cells as soon as they become confluent.
Harvesting LIF conditioned media
When cells are almost confluent (~80-90%) change the media. The following day collect the conditioned media in a 50ml centrifuge tube(s)Spin at 2000 rpm for 5 min to remove any cells. Filter through a 0.2& filter and aliquot into cryovials or parafilmed-centrifuge tubes. Store the stock LIF at -80°C. You can usually add fresh media to these cells and harvest a second batch of LIF the next day. Throw these cells out when you are done as, they have now been confluent for ~2 days. LIF conditioned media CAN be stored short term (a few moths) at -20°C.
Titering LIF
Plate 1 x 104 feeder-free ES cells in each of the wells of a gelatinized 6-well in ES media without LIF. At plating add LIF at 1:100, 1:500, 1:0, 1:5000 and 1:10,000 dilutions from stock. Check for differentiation over a 1 week period. Select the lowest dilution than that which allows any amount of visible differentiation. Usually 1:2500 or 1:5000 is best(If the ES cells are becoming confluent in less than 1 week you may need to plate a smaller number)
Notes: The cells can occasionally be reselected in 0.1mM methotrexate to make sure they are still stably transformed for the LIF transgene. When growing in methotrexate the calf serum must be dialyzed.
Care and Handling of Feeder Layer Cells
STO/SNL cells are a derivative of the standard STO cell line. They are transformed with a LIF producing plasmid (low-level). They grow quite well, and look slightly more vacoulated than regular STO cells. This line was made available to use from the Bradfield lab.
High glucose DMEM (-pyruvate,-glutamine)
Newborn calf serum (heat inactivated 56°C for 30 min)
1X l-glutamine (add fresh every 2 weeks)
1X penicillin/streptomycin
Add all components to DMEM, sterile filter if desired
Thaw frozen vial of cells
Thaw rapidly at 37°C, rinse outside of vial with 70% ethanol
Use a lml pipet to transfer cells to a 15ml centrifuge tube, add 5ml STO/SNL media
Spin down cells for 2 minutes at 1000 rpm
Aspirate off media and resuspend cell pellet in fresh media, plate
Growing cells
Grow cells in 10 cm tissue culture dishes
Change media every 2 days
Split cells 1:10 as soon as they become confluent (usually ~ 4 days)
Splitting cells
Aspirate all media from plate and add ~5 ml. incomplete PBS. aspirate. Repeat 2X. Add 3ml. trypsin-EDTA.
Put plate in 37°C incubator for 8-10 minutes
Tap plate to check that cells are loosened
Add T-E solution to an equal volume of media in a centrifuge tube.
Pipette back and forth a few times to insure a single-cell suspension. Spin 2 min at 2000 rpm.
Aspirate media, resuspend cell pellet in fresh media and plate. Usually cells are plated 1:8.
STO/SNL Cells
High glucose DMEM (-pyruvate,-glutamine)
Newborn calf serum (heat inactivated 56°C for 30 min)
1X l-glutamine (add fresh every 2 weeks)
1X penicillin/streptomycin
Add all components to DMEM, sterile filter if desired
Thaw frozen vial of cells
Thaw rapidly at 37°C, rinse outside of vial with 70% ethanol
Use a lml pipet to transfer cells to a 15ml centrifuge tube, add 5ml STO/SNL media
Spin down cells for 2 minutes at 1000 rpm
Aspirate off media and resuspend cell pellet in fresh media, plate
Growing STO/SNL
Grow cells in 10 cm tissue culture dishes
Change media every 2 days
Split cells 1:10 as soon as they become confluent (usually ~ 4 days)
Splitting STO/SNL cells
Aspirate all media from plate and add ~2ml 1X trypsin, swirl so all cells are covered
Put plate in 37°C incubator for 5 minutes
Tap plate to check that cells are loosened
Add 3ml media to plate, pipet up and down several times to achieve a single-cell suspension
Transfer cells to a 15ml centrifuge tube, spin 2 min at 1000 rpm
Aspirate media, resuspend cell pellet in fresh media and plate
Freezing STO/SNL cells
Trypsinize either regular or MitoC-treated STO/SNL cells as aboveAfter centrifuging resuspend cells in freezing media (STO/SNL media with 20% NCS and 10% DMSO)Volume should be approximately lml per plate of cells, aliquot to cryovials adding lml per vialFreeze vials immediately at -20°C for ~2 hr, then at -80°C overnight, transfer to liquid N2 the next morning.
Growth-Arrest for Feeder Layers
Split cells as described above. However, instead of plating, transfer the cells (in media) to a sterile centrfuge tube. Parafilm the tube, and place the cells in a small ice-bucket. Bring the cells down to the gammacell (you MUST be certified to use the gammacell. See Rex for information about certification). Place the cells in the sample cavity (don’t forget to sign in) and irradiate for 40 minutes. Wipe off the tube with 70% ethanol, and place the tube back in the hood. Plate the cells, using 1 full plate of cells for 1 plate of feeders. One 10cm2 plate makes enogh feeder cells for another 10cm2 plate, 2 24-well plates, or 2 96-well plates. Check the cells and change media (to ES cell media) the next day. It is not unusual to see some death, but normally the cells spread out enough that this does not cause problems. You want the cells to cover the entire plate, confluency is preffered. You do not wish plated ES cells to be exposed to open surfaces on the plate (although they generally only stick to the feeder cells anyway).
Plating MitoC treated feeder cells
A confluent 10 cm plate of STO/SNL cells will make: 2 10 cm plates, 2 6-well plates, 1 24-well of feedersAdd the appropriate MitoC-treated cells to gelatinized wells in STO/SNL mediaAllow 12 hours for cells to attach before using feeders. If after 12 hours feeder cells appear too thin more may be added.
请您对这篇文章发表看法
相关技术资料
关于生物在线
联系人：胡先生Email：电话：021-}